Voip Avaya Ip Guide 3 0
Voip Avaya Ip Guide 3 0
Voip Avaya Ip Guide 3 0
ABSTRACT
This document gives implementation guidelines for the Avaya MultiVantage™ Communications
Applications. Configurations and recommendations are given for various Avaya Media Servers and
Gateways, as well as Avaya 4600 Series IP Telephones. This document also provides information on
virtual local area networks (VLANs), and guidelines for configuring Avaya and Cisco networking
equipment in VoIP applications.
The intent of this document is to provide training on IP telephony, and to give guidelines for
implementing Avaya solutions. It is intended to supplement the product documentation, not replace it.
This document covers the Avaya Communication Manager 2.1 and 3.0, and the Avaya 4600 Series IP
Telephone 1.8 and later, with limited information regarding previous and future versions.
July 2005
COMPAS ID 95180
The instructions and tests in this document regarding Cisco products and features are best-effort attempts
at summarizing and testing the information and advertised features that are openly available at
www.cisco.com. Although all reasonable efforts have been made to provide accurate information
regarding Cisco products and features, Avaya makes no claim of complete accuracy and shall not be
liable for adverse outcomes resulting from discrepancies. It is the user’s responsibility to verify and test
all information in this document related to Cisco products, and the user accepts full responsibility for all
resulting outcomes.
Avaya and the Avaya Logo are trademarks of Avaya Inc. or Avaya ECS Ltd., a wholly owned subsidiary
of Avaya Inc. and may be registered in the US and other jurisdictions. All trademarks identified by ® and
™ are registered trademarks or trademarks, respectively, of Avaya Inc. All other registered trademarks or
trademarks are property of their respective owners.
Cost saving is one factor. By eliminating a separate circuit-switched voice network, businesses avoid the
expenses of buying, maintaining and administering two networks. They may also reduce toll charges by
sending long distance voice traffic over the enterprise network, rather than the public switched telephone
network.
Another benefit is the potential to more tightly integrate data and voice applications. Because they use
open programming standards, Avaya MultiVantage™ products make it easier for developers to create,
and for companies to implement, applications that combine the power of voice and data in such areas as
customer relationship management (CRM) and unified communications. A converged multi-service
network can make such applications available to every employee.
These benefits do not come free, however. Voice and data communications place distinctly different
demands on the network. Voice and video are real-time communications that require immediate
transmission. Data does not. Performance characteristics that work fine for data can produce entirely
unsatisfactory results for voice or video. So networks that transmit all three must be managed to meet the
disparate requirements of data and voice/video.
Network managers are implementing a range of techniques to help ensure their converged networks meet
performance standards for all three payloads: voice, video and data. These techniques include the
strategic placement of VLANs, and the use of Class of Service (CoS) packet marking and Quality of
Service (QoS) network mechanisms.
For an overview of IP telephony issues and networking requirements, see the “Avaya IP Voice Quality
Network Requirements” white paper.
Professional consulting services are available through the Avaya Communication Solutions and
Integration (CSI) group. One essential function of this professional services group is to provide pre-
deployment network assessments to Avaya customers. This assessment helps to prepare a customer’s
network for IP telephony, and also gives critical network information to Avaya support groups that will
later assist with implementation and troubleshooting. Arrange for this essential service through an Avaya
account team.
Table of Contents
References................................................................................................................................................... 77
This section provides a foundation to build upon for the rest of this document. Voice over IP (VoIP)
terminology and Avaya products and architectures are introduced here.
Servers
Most of the intelligence in a voice system is in the call server. From servicing a simple call to making
complex decisions associated with large contact centers, the call server is the primary component of an IP
telephony system. Avaya Communication Manager is the call processing software that runs on Avaya
server platforms.
The following are some common terms for a call server. Some are generic and some are specified by a
protocol, but all are generally used throughout the industry.
Gateways
A gateway terminates and converts various media types, such as analog, TDM, and IP. A gateway is
required so that, for example, an IP phone can communicate with an analog phone on the same telephony
system, as well as with an external caller across a TDM trunk.
The following are some common terms for a gateway, and they are generally used throughout the
industry.
A gateway requires a call server to function, and some common Avaya server-gateway architectures are
illustrated later.
Stations
There are several technical terms for what most would call a telephone, and some that are generally used
throughout the industry are listed below.
- Endpoint – H.323 general term that includes IP phones and other endpoints
- Terminal – H.323 specific term to mean primarily IP phones (also an Avaya term)
- Station – Avaya term, and possibly a generic term
- Set – Avaya term, and possibly a generic term
Avaya gateways have port boards or media modules that terminate various types of stations.
Avaya gateways have port boards or media modules that terminate various types of trunks, including IP
trunks.
The following figures illustrate some common Avaya server-gateway architectures in succession, from
established to most recent technologies. Also included in the diagrams are the protocols used to
communicate between the various devices.
EPN
TDM bus TDM bus
STN
to P
- The single- (SCC1) and multi-carrier cabinets (MCC1) are called port networks (Avaya term) or
media gateways (VoIP term). They house port boards, which, among other things, terminate stations
and trunks. (These port boards are not the focus of this document.)
- The DEFINITY® call servers are the processor boards inserted into the processor port network
(PPN).
- The other cabinets, without processors, are called expansion port networks (EPN) and are controlled
by the DEFINITY servers in the PPN.
- The port networks are connected together via a port network connectivity (PNC) solution, which can
be TDM-based (Center Stage PNC) or ATM-based (ATM PNC). Both bearer (audio) and port
network control are carried across the PNC solutions.
- Control Channel Message Set (CCMS) is the Avaya proprietary protocol used by the DEFINITY
servers to control the port networks (cabinets and port boards).
IP IP
H.225 - RAS &
IP IP Q.931 signaling
H.225
EPN PPN EPN
CCMS from processor
to port boards across
backplane RTP
Procr
Procr
C-LAN audio
IP Net C-LAN
MedPro
MedPro
RTP EPN Enterprise
TDM bus TDM bus IP Network
STN
CCMS and bearer Center Stage
Analog Analog
to P
to P
over TDM or ATM or ATM PNC
- IP-enabled DEFINITY System is the same architecture as before, but with IP port boards added.
- The Control-LAN (C-LAN) board is the call servers’ interface into the IP network for call signaling.
H.225, which is a component of H.323, is the protocol used for call signaling. H.225 itself has two
components: RAS for endpoint registration, and Q.931 for call signaling.
- The IP Media Processor (MedPro) board is the IP termination point for audio. As of Communication
Manager 3.0 there is a higher capacity version of the MedPro board called IP Media Resource 320
(MR320). The MR320 supports up to 320 calls, based on licensing, as opposed to a fixed max of 64
calls for the MedPro. These boards perform the conversion between TDM and IP. The audio payload
is encapsulated in RTP, then UDP, then IP.
IPSI
IPSI
IPSI
IPSI
RTP
C-LAN audio
IP Net C-LAN
MedPro
MedPro
RTP PN Enterprise
IPSI
IPSI
TDM bus TDM bus IP Network
STN
CCMS and bearer Center Stage
Analog Analog
to P
to P
over TDM or ATM or ATM PNC
Figure 3: Multi-Connect
- Multi-Connect is the same underlying DEFINITY architecture, except that the processor boards are
replaced with much more powerful Avaya S8700 or S8710 Media Servers.
- Port networks get IP Server Interface (IPSI) boards to communicate with the S87xx call servers.
- CCMS exchanges between the call servers and port networks now take place over the control IP
network.
- Not all port networks require IPSI boards, because Center Stage PNC and ATM PNC are still present
to connect the port networks.
The Avaya S8500 Media Server is the simplex equivalent of the S87xx server pair. The S8500 gives the
same call processing capability without the redundancy and added reliability of duplicated servers. The
S8500 can be substituted in place of the S87xx servers in any IP-Connect configuration shown below.
G650
MedPro/MR320 boards, as
to P
Analog DCP
well as the other traditional
port boards used in the MCC1
Figure 5: IP-Connect
and SCC1.
S8300/G700/G350/G250
- The Avaya G700, G350, and G250 (not shown) Media Small/Medium Enterprise
Gateways are based on the H.248 protocol.
- One primary difference between these gateways is capacity. H.225
(Refer to current product offerings for exact specifications.) g700 with
- All gateways have built-in Ethernet switches. The G700 IP IP s8300 ICC
supports IP routing and IP WAN connectivity with an VoIP mod
DCP Analog
PS
backup H.225
s8700 s8700 H.225 - RAS &
Q.931 signaling
g700 with
IP IP s8300 LSP
L2 switch L2 switch Control IP IP VoIP mod
IP Network backup
H.248
CCMS H.225
PN PN IP Net g700 with
IPSI
IPSI
IPSI
IPSI
RTP VoIP mod
l
audio N ntro
WA w ay co
C-LAN ate
ia g g350 with
48 med VoIP mod
MedPro H .2
PN
IPSI
IPSI
C-LAN
g700 with
RTP VoIP mod
SCC MCC MCC Enterprise IP IP DCP mod
IP Network Analog mod
CCMS and bearer T1/E1 mod
TN
or ATM PNC
N
DCP Analog
ST
P
to
Figure 7: Multi-Connect with remote G700/G350/G250s
- Remote gateways and stations are controlled by the S87xx servers via the C-LAN boards.
- The remote S8300 is in local survivable processor (LSP) mode to take over as call server if
connectivity to the S87xx servers is lost.
backup H.225
s8700 s8700
G650
to P
N
Analog DCP DCP Analog
ST
P
to
Figure 8: IP-Connect with remote G700/G350/G250s
- Remote gateways and stations are controlled by the S87xx servers via the C-LAN boards.
- The remote S8300 is in local survivable processor (LSP) mode to take over as call server if
connectivity to the S87xx servers is lost.
DCP
QSIG or DCS
PSTN
H.323 (Q.931)
H.225 I P IP
S8700
Public Switch
Q.931
PRI SS7
G650
Public Switch
QSIG or DCS
Inband
Q.931 Q.931
H.225 T1 OR
PRI PRI
IP
Public Switch
Loop Start Inband
T1
OR
QSIG
Q.931
Analog
Vendor X PBX PRI
DEFINITY System
Figure 9: Trunks
QSIG is the standard signaling protocol that provides the feature-richness expected in enterprises.
Generally speaking, traditional telephony systems support a broad range of QSIG features, while pure IP
telephony systems support a very limited range. Due to the history and leadership of Avaya in traditional
telephony, all Avaya call servers – whether traditional, IP-enabled, or pure IP – support virtually the same
broad range of QSIG features.
The following figure illustrates the protocol stacks relevant to VoIP. The contents of the upper-layer
protocol messages are important to those who develop VoIP applications. However, those who
implement these applications are typically not concerned with decoding the upper-layer messages.
Instead, they are concerned primarily with the transport mechanism – TCP and UDP ports – so that they
can verify and filter these message exchanges.
L3 - IP L3 L3
- H.323 is the prevalent VoIP protocol suite. It is used for signaling from gatekeeper to terminals
(stations), and gatekeeper to gatekeeper (trunks).
- H.225 is the endpoint registration (RAS) and call signaling (Q.931) component of H.323.
- H.225 call signaling messages are transported via TCP with port 1720 on the gatekeeper side.
- H.225 registration messages (commonly referred to simply as RAS message) are sent via
UDP with port 1719 on the gatekeeper side.
- H.245 is the multimedia control component of H.323.
- Audio is digitally encoded prior to transmission and decoded after transmission using a coder/decoder
(codec).
- G.711 is the fundamental codec based on traditional pulse-code modulation (PCM), and it is
generally recommended for LAN transport.
- G.729 is a compressed codec, and it is generally recommended for transport over limited-
bandwidth WAN links.
- Encoded audio is encapsulated in RTP (real-time protocol), then UDP, then IP.
- RTP has fields such as Sequence Number and Timestamp that are designed for the transport and
management of real-time applications.
- On Avaya solutions the UDP ports used to transport RTP streams are configured on the call
server.
- Most protocol analyzers can identify RTP packets, making it easy to verify that audio streams are
being sent between endpoints.
- H.248 is a protocol for media gateway control. It is transported via TCP with port 2945 (1039 if
encrypted) on the media gateway controller side.
- CCMS is an Avaya proprietary protocol for port network control (same as media gateway control). It
is transported via TCP with port 5010 on the port network (IPSI board) side.
This section gives general guidelines and addresses several issues related to IP networks (LAN/WAN)
and device configurations.
Because of the time-sensitive nature of VoIP applications, VoIP should be implemented on an entirely
switched network. Ethernet collisions – a major contributor to delay and jitter – are virtually eliminated
on switched networks. Additionally, VoIP endpoints should be placed on separate subnets or VLANs
(separated from other non-VoIP hosts), with preferably no more than ~500 hosts per VLAN. This
provides a cleaner design where VoIP hosts are not subjected to broadcasts from other hosts, and where
troubleshooting is simplified. This also provides a routed boundary between the VoIP segments and the
rest of the enterprise network, where restrictions can be placed to prevent unwanted traffic from crossing
the boundary. When a PC is attached to an IP telephone, even if they are on separate VLANs, both PC
and phone traffic (including broadcasts) traverse the same uplink to the Ethernet switch. In such a case
the uplink should be a 100M link, and the recommended subnet/VLAN size is no larger than ~250 hosts
each for the voice and data VLANs.
High broadcast levels are particularly disruptive to real-time applications like VoIP. Avaya media servers
and gateways and IP telephones utilize very low amounts of broadcast traffic to operate. Therefore, a
subnet/VLAN with only these Avaya hosts has a very low broadcast level. There are two cases where
Avaya hosts can be subjected to high levels of broadcasts: 1) Avaya hosts and other broadcast-intensive
hosts share a subnet/VLAN; and 2) broadcast-intensive PCs are attached to Avaya IP phones. Case 1 is
one of the reasons for the recommendation to use separate voice subnets/VLANs. Case 2 is sometimes
unavoidable, and the result is that broadcasts used by the PC must pass through the phone, even if the
phone and PC are on different VLANs. For this reason Avaya IP phones are designed to be very resilient
against broadcasts, with lab tests showing the phones operating satisfactorily even with 3,000 to 10,000
broadcasts per second, depending on the model. Nevertheless, to provide acceptable user experience and
audio quality, high-broadcast environments are very strongly discouraged. The recommended maximum
broadcast rate is 500 per second, and the absolute maximum is 1000 per second.
If VoIP hosts must share a subnet with non-VoIP hosts (not recommended), they should be placed on a
subnet/VLAN of ~250 hosts or less with as low a broadcast rate as possible. Use 100M links, take
measures not to exceed the recommended maximum broadcast rate (500/s), and do not exceed the
absolute maximum broadcast rate (1000/s).
Ethernet Switches
The following recommendations apply to Ethernet switches to optimize operation with Avaya IP
telephones and other Avaya VoIP endpoints, such as IP boards. They are meant to provide the simplest
configuration by removing unnecessary features.
Speed/Duplex
One major issue with Ethernet connectivity is proper configuration of speed and duplex. There is a
significant amount of misunderstanding in the industry as a whole regarding the auto-negotiation
standard. The speed can be sensed, but the duplex setting is negotiated. This means that if a device with
fixed speed and duplex is connected to a device in auto-negotiation mode, the auto-negotiating device can
sense the other device’s speed and match it. But the auto-negotiating device cannot sense the other
device’s duplex setting; the duplex setting is negotiated. Therefore, the auto-negotiating device always
goes to half duplex in this scenario. The following table is provided as a quick reference for how speed
and duplex settings are determined and typically configured. It is imperative that the speed and duplex
settings be configured properly.
Layer 1 (L1) errors such as runts, CRC errors, FCS errors, and alignment errors often accompany a
duplex mismatch. If these errors exist and continue to increment, there is probably a duplex mismatch or
cabling problem or some other physical layer problem. The show port <mod/port> command on
Catalyst switches gives this information. The Avaya P550 commands are show port status <mod/port>,
show port counters <mod/port>, and show ethernet counters <mod/port>. The Avaya P330 switch
command is show rmon statistics <mod/port>.
Calculation
Many VoIP bandwidth calculation tools are available, as well as pre-calculated tables. Rather than
presenting a table, the following information is provided to help the administrator make an informed
bandwidth calculation. The per-call rates for G.711 and G.729 are provided under the “Ethernet
Overhead” and “WAN Overhead” headings below, and all calculations are for the recommended voice
packet size of 20ms.
- Voice payload and codec selection – The G.711 codec payload rate is 64000bps. Since the audio is
encapsulated in 10-ms frames, and there are 100 of these frames in a second (100 * 10ms = 1s), each
frame contains 640 bits (64000 / 100) or 80 bytes of voice payload. The G.729 codec payload rate is
8000bps, which equates to 80 bits or 10 bytes per 10-ms frame.
Voice Payload 1 frame – 10ms 2 frames – 20ms 3 frames – 30ms 4 frames – 40ms
G.711 80 B 160 B 240 B 320 B
G.729 10 B 20 B 30 B 40 B
Table 2: Voice payload vs. number of frames
- Packet size and packet rate – Because the voice payload rate must remain constant, the number of
voice frames per packet (packet size) determines the packet rate. As the number of frames per packet
increases, the number of packets per second decreases to maintain a steady rate of 100 voice frames
per second (64000bps or 8000bps).
To this point the calculation is simple. Add up the voice payload and overhead per packet, and multiply
by the number of packets per second. Here are the calculations for a G.711 and a G.729 call, both using
20-ms packets. (Remember that there are 8 bits per byte.)
The calculations above do not include the L2 encapsulation overhead. L2 overhead must be added to the
bandwidth calculation, and this varies with the protocol being used (Ethernet, PPP, HDLC, ATM, Frame
Relay, etc).
WAN Overhead
The WAN overhead is calculated just like the Ethernet overhead, by determining the size of the L2
encapsulation and figuring it into the calculation. L2 headers and trailers vary in size with the protocol
being used, but they are typically much smaller than the Ethernet header and trailer. For example, the
PPP overhead is only 7 bytes. However, to allow for a high margin of error, assume a 14-byte total L2
encapsulation size, which would add an overhead of 5.6kbps (14 * 8 * 50), assuming 20-ms voice
packets. This would result in approximately 86kbps
G.729 20-ms call over PPP = 26.8kbps for G.711 and 30kbps for G.729 over a WAN link.
G.729 30-ms call over PPP = 20.5kbps Significant bandwidth savings are realized by using a
compressed codec (G.729 recommended) across a
G.729 20-ms call over 14-B L2 = 29.6kbps WAN link. Note that in today’s data networks most, if
G.729 30-ms call over 14-B L2 = 22.4kbps not all, WAN links are full duplex.
MTU should not be an issue for VoIP because most interfaces have a default MTU of 1500 bytes.
However, it is possible to intentionally set the MTU to low levels. Even if the MTU is not set to a level
that would fragment VoIP packets, the principle of fragmenting the L3 payload and incurring additional
L3 and L2 overhead applies universally. Changing the MTU requires a thorough understanding of the
traffic traversing the network. A low MTU value, relative to any given IP packet size, always increases
L3 and L2 overhead as illustrated with the VoIP example. Because of this inefficiency, it is generally not
advisable to lower the MTU.
L2 Fragmentation
The second factor involves fragmenting the L2 payload, or the entire IP packet. This process of
fragmenting a single IP packet into multiple L2 frames incurs additional L2 overhead, but no additional
IP overhead. For example, the fixed cell size for ATM is 53 octets (bytes), with 5 octets for ATM
overhead and 48 octets for payload. Without header compression there is no way to get a voice packet to
fit inside one ATM cell. Therefore, the L3 packet (not just the IP payload, but the entire IP packet) is
fragmented and carried inside multiple ATM cells. A 200-byte G.711 IP packet would require five ATM
cells (25 octets of ATM overhead), whereas a 60-byte G.729 IP packet would only require two ATM cells
(10 octets of ATM overhead). Refer to Appendix C for information regarding RTP header compression.
Keep in mind, however, that the same precautions apply to RTP header compression as to QoS (see the
next section on CoS and QoS). The router could pay a significant processor penalty if the compression is
done in software.
Inter-LATA (typically interstate) Frame Relay is also affected by this ATM phenomenon. This is because
most carriers (ATT, Worldcom, Sprint) convert Frame Relay to ATM for the long haul, between the local
central offices. This is done through a process called frame-relay-to-ATM network interworking and
service interworking (FRF.5 and FRF.8). In this process the Frame Relay header is translated to an ATM
header, and the Frame Relay payload is transferred to an ATM cell. Since the Frame Relay payload can
be a variable size but the ATM payload is a fixed size, a single Frame Relay frame can be converted to
multiple ATM cells for the long haul. Therefore, it is beneficial to limit the size of the voice packet even
when the WAN link is Frame Relay.
General
The term “Class of Service” refers to mechanisms that mark traffic in such a way that the traffic can be
differentiated and segregated into various classes. The term “Quality of Service” refers to what the
CoS
802.1p/Q at the Ethernet layer (L2) and DSCP at the IP layer (L3) are two CoS mechanisms that Avaya
products employ. These mechanisms are supported by the IP telephones and most IP port boards. In
addition, the call server can flexibly assign the UDP port range for audio traffic transmitted from the
MedPro/MR320 board or VoIP media module. Although TCP/UDP source and destination ports are not
CoS mechanisms, they are inherently used to identify specific traffic and can be used much like CoS
markings. Other non-CoS methods to identify specific traffic are to key in on source and destination IP
addresses and specific protocols (ie, RTP).
802.1p/Q
The figure below shows the IEEE 802.1Q tag and its insertion point in the Ethernet and 802.3 frames.
Note that in both cases the 802.1Q tag changes the size and format of the comprehensive Ethernet and
802.3 frames. Because of this, many intelligent switches (ones that examine the L2 header and perform
some kind of check against the L2 frame) must be explicitly configured to accept 802.1Q tagged frames.
Otherwise, these switches may reject the tagged frames. The Tag Protocol Identifier (TPID) field has hex
value x8100 (802.1QTagType). This value alerts the switch or host that this is a tagged frame. If the
switch or host does not understand 802.1Q tagging, the TPID field is mistaken for the Type or Length
field, which results in an erroneous condition.
802.2 802.2
802.3 header
LLC SNAP
8 octets total
802.1p/Q
The two other fields of importance are the Priority and Vlan ID (VID) fields. The Priority field is the “p”
in 802.1p/Q and ranges in value from 0 to 7. (“802.1p/Q” is a common term used to indicate that the
Based on the answers to these questions, tagging should be enabled following these two rules.
1. Single-VLAN Ethernet switch port (default scenario).
- On a single-VLAN port there is no need to tag to specify a VLAN, because there is only one
VLAN.
- For priority tagging only, the IEEE 802.1Q standard specifies the use of VID 0. VID 0 means
that the frame belongs on the port’s primary VLAN, which IEEE calls the “port VLAN,” and
Cisco calls the “native VLAN.” Some Ethernet switches do not properly interpret VID 0, in
which case the port/native VID may need to be used, but this is not the standard method.
- For single devices, such as a call server or port board, a simpler alternative is to not tag at all, but
configure the Ethernet switch port as a high-priority port instead. This treats all incoming traffic
on that port as high-priority traffic, based on the configured level.
- For multiple devices on the same VLAN, such as an IP telephone with a PC attached, the high-
priority device (IP telephone) should tag with VID 0 and the desired priority. The low-priority
device (PC) would not tag at all. No tag at all is the same as priority 0 (default).
2. Multi-VLAN Ethernet switch port.
- A multi-VLAN port has a single port/native VLAN and one or more additional tagged VLANs,
with each VLAN pertaining to a different IP subnet.
- In general, do not configure multiple VLANs on a port with only one device attached to it (unless
that device is another Ethernet switch across a trunk link).
- For the attached device that belongs on the port/native VLAN, follow the points given for rule 1
above. Clear frames (untagged frames) are forwarded on the port/native VLAN by default.
- An attached device that belongs on any of the tagged VLANs must tag with that VID and the
desired priority.
- The most common VoIP scenario for a multi-VLAN port is an IP telephone with a PC attached,
where the phone and PC are on different VLANs. In this case the PC would send clear frames,
and the IP telephone should tag with the designated VID and desired priority.
As stated previously, an Ethernet switch must be capable of interpreting the 802.1Q tag, and many must
be explicitly configured to receive it. The use of VID 0 is a special case, because it only specifies a
priority and not a VLAN. Avaya switches accept VID 0 without any special configuration, but some
Ethernet switches do not properly interpret VID 0. And some switches require trunking to be enabled to
accept VID 0, while others do not. The following table shows the results of some testing performed by
Avaya Labs on Cisco switches.
Catalyst 6509 w/ Accepted VID 0 for the native VLAN when 802.1Q trunking was enabled
CatOS 6.1(2) on the port.
Would not accept VID 0 for the native VLAN. Opened a case with Cisco
Catalyst 4000 w/ TAC, and TAC engineer said it was a hardware problem in the 4000. Bug
CatOS 6.3(3) ID is CSCdr06231. Workaround is to enable 802.1Q trunking and tag with
native VID instead of 0.
Catalyst 3500XL w/ Accepted VID 0 for the native VLAN when 802.1Q trunking was disabled
IOS 12.0(5)WC2 on the port.
DSCP
IP Header
32 bits
octet octet octet octet
Header
Version Type of Service Total Length
Length
Flags
Identifier Fragment Offset
(3)
Source Address
Destination Address
The figure above shows the IP header with its 8-bit Type of Service (ToS) field. The ToS field contains
three IP Precedence bits and four Type of Service bits as follows.
000 Routine
001 Priority
Bits 0-2 010 Immediate
011 Flash
IP Precedence 100 Flash Override
101 CRITIC/ECP
110 Internetwork Control
111 Network Control
Bit 3 0 Normal
Delay 1 Low
Bit 4 0 Normal
Throughput 1 High
Bit 5 0 Normal
Reliability 1 High
Bit 6 0 Normal
Monetary Cost 1 Low
Bit 7 Always set to 0
Reserved
Figure 15: Original scheme for IP ToS field
Ideally any DSCP value would map directly to a Precedence/ToS combination of the original scheme.
This is not always the case, however, and it can cause problems on some legacy devices, as explained in
the following paragraph.
On any device, new or old, having a non-zero value in the ToS field cannot hurt if the device is not
configured to examine the ToS field. The problems arise on some legacy devices when the ToS field is
examined, either by default or by enabling QoS. These legacy devices (network and endpoint) may
contain code that only implemented the IP Precedence portion of the original ToS scheme, with the
remaining bits defaulted to zeros. This means that only DSCP values divisible by 8 (XXX000) can map
to the original ToS scheme. For example, if an endpoint is marking with DSCP 40, a legacy network
device can be configured to look for IP Precedence 5, because both values show up as 10100000 in the
ToS field. However, a DSCP of 46 (101110) cannot be mapped to any IP Precedence value alone.
Another hurdle is if the legacy code implemented IP Precedence with only one ToS bit permitted to be set
high. In this case a DSCP of 46 still would not work, because it would require two ToS bits to be set
high. When these mismatches occur, the legacy device may reject the DSCP-marked IP packet or exhibit
some other abnormal behavior. Most newer devices support both DSCP and the original ToS scheme.
QoS on a Router
It is generally more complicated to implement QoS on a router than on an Ethernet switch. Unlike
Ethernet switches, routers typically do not have a fixed number of queues. Instead, routers have various
queuing mechanisms. For example, Cisco routers have standard first-in first-out queuing (FIFO),
weighted fair queuing (WFQ), custom queuing (CQ), priority queuing (PQ), and low latency queuing
(LLQ). LLQ is a combination of priority queuing and class-based weighted fair queuing (CBWFQ), and
it is Cisco’s recommended queuing mechanism for real-time applications such as VoIP. Each queuing
mechanism behaves differently and is configured differently, but following a common sequence. First the
desired traffic must be classified using DSCP, IP address, TCP/UDP port, or protocol. Then the traffic
must be assigned to a queue in one of the queuing mechanisms. Then the queuing mechanism must be
applied to an interface. [2 p.1-7, 3-4, 3-5, 5-2]
Consult Cisco’s documentation for detailed information regarding traffic shaping and LFI, and be
especially careful with LFI. On one hand it reduces the serialization delay, but on the other it increases
the amount of L2 overhead. This is because a single L3 packet that was once transported in a single L2
frame, is now fragmented and transported in multiple L2 frames. Configure the fragment size to be as
large as possible while still allowing for acceptable voice quality.
Instead of implementing LFI, some choose to simply lower the MTU size to reduce serialization delay.
Two possible reasons for this are that LFI may not be supported on a given interface, or that lowering the
MTU is easier to configure. As explained in section 2.2 under the heading “L3 Fragmentation (MTU),”
lowering the MTU (L3 fragmentation) is much less efficient than LFI (L2 fragmentation) because it
incurs additional L3 overhead as well as additional L2 overhead. Lowering the MTU is generally not
advisable and may not provide any added value, because it adds more traffic to the WAN link than LFI.
The added congestion resulting from the increase in traffic may effectively negate any benefit gained
from reducing serialization delay. One should have a thorough understanding of the traffic traversing the
WAN link before changing the MTU.
Because of all these configuration variables, properly implementing QoS on a router is no trivial task.
However, it is on the router where QoS is needed most, because most WAN circuits terminate on routers.
Appendix F contains examples of implementing QoS on Cisco routers. This appendix does not contain
configurations for all the issues discussed in this document, but it gives the reader a place to start.
QoS Guidelines
There is no all-inclusive set of rules regarding the implementation of QoS, because all networks and their
traffic characteristics are unique. It is good practice to baseline the VoIP response (ie, voice quality) on a
network without QoS, and then apply QoS as necessary. Conversely, it is very bad practice to enable
multiple QoS features simultaneously, not knowing what effects, if any, each feature is introducing. If
voice quality is acceptable without QoS, then the simplest design may be a wise choice. If voice quality
is not acceptable, or if QoS is desired for contingencies such as unexpected traffic storms, the best place
One caution to keep in mind about QoS is regarding the processor load on network devices. Simple
routing and switching technologies have been around for many years and have advanced significantly.
Packet forwarding at L2 and L3 is commonly done in hardware (Cisco calls this fast switching [2 p.5-
18], “switching” being used as a generic term here), without heavy processor intervention. When
selection criteria such as QoS and other policies are added to the routing and switching function, it
inherently requires more processing resources from the network device. Many of the newer devices can
handle this additional processing in hardware, resulting in maintained speed without a significant
processor burden. However, to implement QoS, some devices must take a hardware function and move it
to software (Cisco calls this process switching [2 p.5-18]). Process switching not only reduces the speed
of packet forwarding, but it also adds a processor penalty that can be significant. This can result in an
overall performance degradation from the network device, and even device failure.
Each network device must be examined individually to determine if enabling QoS will reduce its overall
effectiveness by moving a hardware function to software, or for any other reason. Since most QoS
policies are implemented on WAN links, the following very general points for Cisco routers are offered to
increase the level of confidence that QoS remains in hardware. Consult Cisco to be sure.
- Newer hardware platforms are required: 2600, 3600, 7200, and 7500.
- Newer interface modules (WIC, VIP, etc.) are required: Consult Cisco to determine which hardware
revision is required for any given module.
- Sufficient memory is required: Device dependent.
- Newer IOS is required: 12.0 or later.
Several things should be examined whenever QoS is enabled on a network device. First, the processor
level on the device should be examined and compared to levels before QoS was enabled. It is likely that
the level will have gone up, but the increase should not be significant. If it is significant, then it is likely
that the QoS process is being done by software. The processor load must remain at a manageable level
(max 50% average, 80% peak). If the processor load is manageable, the VoIP response should be
examined to verify that it has improved under stressed conditions (ie, high congestion) compared to
performance before QoS was implemented. There is no added value in leaving a particular QoS
mechanism enabled if VoIP response has not improved under stressed conditions. If VoIP response has
improved, then the other applications should be checked to verify that their performances have not
degraded to unacceptable levels.
This section gives guidelines for Avaya servers and gateways, and covers most of the IP-telephony-
related configurations. Refer back to section 1 for an overview of IP telephony components and Avaya
architectures.
Avaya Communication Manager is the call processing software that runs on Avaya servers, and it is
configured via the Switch Administration Terminal (SAT) interface. Although the server platforms
themselves may be configured in various ways, SAT is the universal interface for Communication
Manager.
The Avaya Site Administrator (SA) is a client software application used to access the SAT interface on all
Avaya servers. Additionally, on all but the DEFINITY servers, SAT can also be accessed by telnet-ing to
the server.
The S87xx and S8500 are 19-inch rack-mountable Red Hat Linux server platforms. S87xx servers
operate in a redundant pair, whereas the S8500 is a simplex server. Each server is configured and
managed via a variety of web interfaces, with the Maintenance Web Interface being the most
comprehensive. The web interfaces are designed to facilitate all anticipated configuration and
management requirements, and there is little or no need for a customer to access the Linux shell.
In an S87xx pair one of the servers is active and the other is standby. SAT administration is performed on
the active server, and it is automatically carried over to the standby server. Either of the servers could be
active or standby at any given time, and there are different ways to determine which is active. If the two
servers are on the same subnet there is a virtual IP address, which is labeled the active server address in
the Configure Server – Configure Interfaces screen of the Maintenance Web Interface. Whichever
server is active takes control of the active server address, and telnet-ing or browsing or pointing Avaya
SA to that address accesses the active server. If the two S8700 servers are not on the same subnet (server
separation), there is no virtual active server address. The Status Summary web screen shows the status
of the servers.
The S8700 SAT interface may be accessed using Avaya SA or by telnet-ing to port 5023: telnet <active
server address> 5023. This could also be done by telnet-ing to the active server and typing sat from the
Linux shell. The standby server does not permit access to SAT.
SAT access to the S8500 is similar to that of the S87xx server pair, except that there is only one server.
S87xx/S8500 Speed/Duplex
Speed and duplex for the various S87xx/S8500 Ethernet interfaces are configured using the Configure
Server – Configure Interfaces web admin screen. It is critical to configure the speed and duplex
correctly on the server interfaces used to communicate with the IPSI boards. A speed/duplex mismatch
between these interfaces and the Ethernet switch causes severe call processing problems.
The web admin screen has a pull-down menu for the various speed/duplex settings. This pull-down menu
does not indicate the current configuration, but only the available options. A “current speed” description
next to this pull-down menu indicates the current speed and duplex, but it does not indicate whether these
settings were manually configured or reached via auto-negotiation. Follow these steps to properly
configure the speed and duplex.
On an IP-Connect system the port network control traffic traverses the enterprise IP network, which
services various classes of traffic. If QoS is desired and properly configured on this network, it may be
necessary to have the S87xx/S8500 server(s) tag/mark the port network control traffic. This is only
required on the interfaces that communicate with IPSI boards, as they are the only ones that participate in
real-time IP telephony. Traffic is tagged/marked from these interfaces on a per destination basis for each
IPSI board, as administered on the SAT ipserver-interface form (see section 3.4, heading “IP Server
Interface Board”). For the 802.1p priority from the SAT form to be applied to the S87xx/S8500 server(s),
L2 tagging must be enabled on the appropriate server interfaces via the Configure Server – Configure
Interfaces web admin screen. The interfaces that communicate with IPSI boards have this option, and
the others do not. The VLAN ID is always 0 for the S87xx/S8500 servers (follow the instructions in
section 2.3, heading “Rules for 802.1p/Q Tagging”).
The S8300 is a Red Hat Linux server platform, similar to the S87xx/S8500, but on a compact media
module that fits into a G700/G350/G250 gateway (always in media module slot 1). The S8300 is similar
to the S87xx/S8500 in many ways. It is configured and managed via the same web interfaces, and, as
with the other servers, there is little or no need for a customer to access the Linux shell.
In a G700 the S8300 must have an IP address on the same IP subnet as the MGP, with the same mask and
default gateway (see G700 section below). This is because all media module slots in a G700 inherit the
VLAN of the MGP, and therefore all VoIP media modules and the S8300 must be on the same IP subnet
as the MGP. In a G350/G250 a VLAN must be designated as the ICC VLAN, and the S8300 must have
an IP address on the IP subnet pertaining to that VLAN (see G350 section below).
An S8300 server can be in one of two modes: internal call controller (ICC) or local survivable processor
(LSP). In ICC mode the S8300 is a standalone call server. In LSP mode it is a backup to the primary call
The S8300 SAT interface may be accessed using Avaya SA or by telnet-ing to port 5023: telnet <S8300
address> 5023. This could also be done by telnet-ing to the S8300 and typing sat from the Linux shell.
S8300 ICC permits SAT configuration (changes and displays), but S8300 LSP does not (displays only)
because it receives its Avaya Communication Manager translations from the primary server.
The S8300 connects to the G700/G350/G250 via a backplane 100M Ethernet interface, which is not
configurable.
Three components of the P330 should be configured: the inband management interface, the default
route, and the switch itself. The inband management interface is displayed and configured using the
commands show interface inband and set interface inband, respectively. The inband interface requires
a VLAN, an IP address, and a mask. The VLAN can be any of the VLANs active on the P330, and the IP
address and mask must correspond to the IP subnet associated with that VLAN.
Once configured, the inband interface should be thought of as a host attached to the P330. This may seem
non-intuitive, because the inband interface is the P330 and the way to administer the P330 remotely.
However, like most L2 switch management interfaces, the inband interface is associated with a specific
VLAN. As such, it is accessed just like any other host attached to the switch on a given VLAN – either
directly from another host on the same VLAN/subnet, or by routing to it from a host on a different
VLAN/subnet. Many mistakenly think that any host attached to the P330 should be able to access the
inband interface directly, and this is not necessarily true. Hosts on different VLANs/subnets must route
to the inband management interface via a L3 router.
Like any other IP host, the inband interface needs a default route if it is to route off of its VLAN/subnet.
The default route for the inband interface is displayed and configured using the commands show ip route
and set ip route, respectively. If there is more than one router on the inband VLAN/subnet, the inband
interface may have additional routes based on destination subnets or hosts. These are displayed and
configured using the same commands.
Finally, the P330 L2 switch itself has various configuration parameters, such as Spanning Tree, VLANs,
trunking, and speed/duplex. These are configured just like on the P330 switch (see appendix E).
Like the P330 inband management interface, the MGP should be thought of as a host on the P330 L2
switch. The command session mgp from the P330 CLI puts the user into the MGP CLI. The MGP
requires a VLAN, IP address, and mask. These are displayed and configured using the MGP CLI
commands show interface mgp and set interface mgp (type configure to enter configuration mode for
the set commands). The MGP may be on the same VLAN as the inband interface, or on a different
VLAN. If on a different VLAN, a L3 router is required to route between the two VLANs. Like the
inband interface, the MGP also needs at least a default route to route off of its VLAN/subnet. The MGP
CLI commands are show ip route mgp and set ip route to display and configure MGP routes.
Each VoIP media module also requires an IP address using the set interface voip v# command. The
VoIP modules inherit the VLAN, mask, and configured routes of the MGP, so there is no need to
explicitly configure them for each VoIP module. The internal VoIP module is voip v0. An external VoIP
module would be voip v1 or voip v2 or voip v3 or voip v4, depending on which slot it is in. show mm
shows all the media modules and their slot numbers.
Device and uplink management are key factors. If several G700 gateways are co-located in the same
rack, it makes practical sense to use the Octaplane stacking feature. This allows the P330 components of
all the G700s to be managed via a single inband interface. But more importantly, it eliminates the need
for each G700 to be uplinked to the next network device individually.
When determining whether or not a single G700 should be added to an existing Octaplane stack of P330
switches, the relative importance of the G700 to the other devices is another factor. A G700’s primary
role is that of IP telephony, specifically media conversion. A P330 switch’s primary role is that of L2
Bandwidth is another key factor for using or not using the Octaplane stack. The G700 components (P330
inband, MGP, VoIP modules, S8300) require a certain amount of bandwidth to communicate off the
chassis. Each VoIP module consumes a maximum of approximately 6Mbps to service 64 G.711 calls
using 20-ms packets. With up to five VoIP modules on a single G700, the maximum bandwidth
consumption is approximately 30Mbps. Other than firmware and translation downloads the bandwidth
consumed by the other components is negligible. Therefore, a single 100M uplink from EXT1 or EXT2
to another Ethernet switch is sufficient for the G700 itself.
The added bandwidth of the Octaplane stack might be required when the 16-port X330 Ethernet
expansion module is used in the G700, and the hosts attached to that module communicate mostly to
other hosts not on the G700. If the hosts on the expansion module are IP telephones, a 100M uplink is
sufficient. But if PCs are attached to the phones and the PCs frequently communicate off the G700, a
100M uplink may not be sufficient.
Two significant differences between the G350 and G700 are capacity and architecture. The G350
supports much fewer users (~40 max) than the G700 (~450max). As such, the G350’s internal VoIP
module has only 32 audio resources, as opposed to 64 in the G700’s internal VoIP module, and in the
external VoIP media module and MedPro board. The G350 also cannot presently accept external VoIP
modules.
The primary architectural difference between the G350 and G700 is that the G350 is an integrated
platform. The L2 switch, MGP, and internal VoIP module all share the same processing engine and same
IP address. In addition, a L3 router is integrated into the G350, whereas the G700 can accept a L3 router
as an expansion module. The resulting G350 CLI has two components. The L2 switch and MGP
commands are practically the same as on the G700, using set commands similar to the P330 switch and
Cisco’s CatOS. The L3 router commands are practically the same as on the X330WAN and P330R
routers, using commands similar to Cisco’s IOS. The G350 utilizes both command sets in a single CLI.
The G350 can have several physical and logical interfaces, and there is no single inband interface as on
the G700’s P330 switch component. One of the G350’s IP interfaces must be designated the primary
management interface (PMI). There is a default designation, but it can be changed by inserting the
command pmi under the desired interface. The PMI, among other things, is the interface used by the
MGP and internal VoIP module. This means that H.248 media gateway signaling and RTP audio are
sourced by and terminated on the PMI. When an S8300 media server is inserted into a G350, one of the
VLANs on the G350 must be designated the ICC VLAN. This is done by inserting the command icc-
vlan under the desired VLAN interface.
The remaining G350/G700 similarities and differences are as follows. Refer to the G700 sections above
for more details on each subject.
Like the G700 and G350, the G250 is capable of housing an S8300 server in ICC or LSP mode. In
addition, unlike the other two gateways, the G250 can act as a survivable H.323 gatekeeper. This feature
is called Standard Local Survivability (SLS), and it allows the G250 to be a call server with very limited
features when communication to the primary call server is lost. For simplicity, SLS can be thought of as
an integrated LSP with very limited features. The details of SLS and its configuration are not covered in
this document.
When connecting a gateway to another Ethernet switch, the uplink between the two switches should be
fixed at 100M/full-duplex (see section 2.1, heading “Speed/Duplex”). Furthermore, if 802.1p/Q tagging
information – VLAND ID and/or L2 priority – is to be passed across the uplink, both switches must have
802.1Q trunking enabled with matching VLANs on the connected ports.
For survivability reasons at remote offices, the LSP, media gateway, and local IP phones should all be on
the same voice VLAN/subnet. If the LSP, media gateway, and IP phones are on different subnets, they
depend on a router or routing process to function. This is not desirable, especially at a branch office
where there is typically only one router. The IP telephony system should be able to function even if that
router or routing process fails. In larger remote offices it may not be feasible to put all VoIP endpoints on
a single subnet, but the principle of minimizing dependencies still holds.
For remote offices where the WAN link terminates on the gateway, whether on the X330WAN router or
natively on the G350/G250 itself, the DSCP values for audio and signaling must be 46 and 34
respectively. The X330WAN router in “voip-queue” mode and the G350/G250 gateways are optimized
to use these values for QoS on the WAN link. These values can be configured locally via the set qos
bearer/signal commands, or they can be downloaded from Communication Manager. On
Communication Manager these values are configured on the SAT ip-network-region form for the region
to which the gateway is assigned. A gateway is configured to use the Communication Manager values by
executing set qos control remote on the MGP/gateway CLI. The CLI command show qos-rtcp displays
the locally set and remotely downloaded values, as well as which values are in use.
The G650/G600, MCC1, and SCC1 are non-H.248 media gateways. They are controlled via the Avaya
CCMS protocol, unlike the G700/G350/G250 gateways which are controlled via the H.248 protocol. The
CCMS-based gateways are better known as port networks, and they share the same port boards. The most
significant boards related to IP telephony are the C-LAN (TN799DP), MedPro (TN2302AP), and MR320
(TN2602AP) boards. Boards with these specific codes are required for Communication Manager;
previous board revisions cannot be used.
The default speed/duplex setting on the MedPro/MR320 board is auto-negotiate. The default
speed/duplex setting on the TN799DP C-LAN board is 10/half, to make it backwards compatible with the
previous TN799C board, which could only do 10/half. When a C-LAN or MedPro/MR320 is inserted
into one of the port networks, the board receives its speed/duplex programming from Communication
Manager, per the appropriate form. If for any reason a board loses this programming, it reverts back to
the board’s default.
If there is poor audio quality on calls going through a particular MedPro/MR320 board, follow these steps
to determine if a speed/duplex mismatch between the MedPro/MR320 and the Ethernet switch is the
cause.
- Check both the board (get ethernet-options <slot #>) and the Ethernet switch port and verify that
they are set to the same speed/duplex or have auto-negotiated to the same speed/duplex.
- Check for L1 errors as instructed in section 2.1 under the “Speed/Duplex” heading.
- Send a continuous ping (ping -t) to the MedPro/MR320 from a Windows machine. If the pings
intermittently fail and the failures coincide with periods of poor audio quality, then there is probably a
speed/duplex problem between the board and the Ethernet switch.
The Cisco white paper “Troubleshooting Cisco Catalyst Switches to Network Interface Card (NIC)
Compatibility Issues [4 p.6]” describes the flapping problem mentioned above and offers a suggestion to
adjust the jitter tolerance (not related to audio jitter) on Cisco switches. The CatOS global command
(which is hidden) is set option debounce enable (disable to undo). This command increases the jitter
tolerance to 3.1 nsec from the 1.4-nsec default. The IOS interface command is carrier-delay 4 (no
carrier-delay to undo). This adjusts the carrier transition delay to 4 seconds. If these commands do not
correct or improve the flapping condition, put the switch back to its original state and try operating at
10/half until the problem can be resolved.
If IP Control is ‘y’ the board is acting as an IPSI; otherwise (‘n’) it is acting as a tone clock. The ‘n’
option is primarily used for migrating a non-IPSI port network to an IPSI port network. Ignore
Connectivity in Server Arbitration has to do with whether or not connectivity to this IPSI is factored into
the decision to interchange S87xx servers. In most cases this is set to ‘n’, but in rare cases it could be set
to ‘y’ for IPSIs in remote locations with poor network connectivity back to the servers. The intent would
be to avoid server interchanges caused by frequent and inconsistent loss of communication to this IPSI.
Location is the board slot #. Host is the board’s static IP address if configured manually, or the hostname
if the address was obtained via DHCP. DHCP ID is the hostname. Socket Encryption, if the parameter is
present, allows the control link between the IPSI and call server to be encrypted. When QoS is enabled
the 802.1p and DiffServ parameters contain the values to be applied to the call server when
communicating with this IPSI board (values are not applied to the IPSI board itself).
The IPSI’s speed/duplex and L2/L3 priority values are configured on the board itself, instead of via SAT
forms. From the IPSI board type ipsilogin at the [IPSI]: prompt, and enter the login name and password
to access the [IPADMIN]: prompt. The commands to display and configure the control port speed and
duplex are show port 1, set port negotiation 1, set port speed 1, and set port duplex 1. The commands
to display and configure the L2 and L3 priority values are show qos, set vlan tag, set vlan priority, and
set diffserv. Be sure to understand what these values do before setting them (see all of section 2.3,
particularly the heading “Rules for 802.1p/Q Tagging”).
The SAT interface has various “forms” that are used to configure specific features. This section covers
the forms used to configure general IP telephony. Most of the forms have a display option to view the
current configurations, and a change option to change them. Some also have a list option to view, for
example, a broad list of stations without seeing in detail how each station is configured.
ethernet-options
As of Avaya Communication Manager 2.0 each IP board’s speed and duplex settings are configured using
the ip-interface form. The ethernet-options form has the list and get options to verify actual
speed/duplex settings against configured settings for all boards and individual boards respectively. With
each new system or IP board installation, one standard procedure should be to apply matching
speed/duplex settings to each IP board and its corresponding Ethernet switch port.
node-names ip
Options are change and display. This form is used to define arbitrary names and associate an IP address
with each name. For example, the name “c-lan_80” could be defined to describe a C-LAN board on the
80 subnet with address 192.168.80.10, and the name “medpro_80” could be defined to describe a MedPro
board on the 80 subnet with address 192.168.80.11.
ip-interface
Options are change, display, and list. This form is used to configure individual IP boards. The first step
is to associate a board Type and Slot # to a previously defined Node Name, and to give the board a Subnet
Mask and default Gateway and assign it to a Network Region. For example, the board type C-LAN in
slot 01A05 can be associated with the node name “c-lan_80” defined earlier. This assigns the IP address
192.168.80.10 to the C-LAN board in slot 01A05. Then the board can be given the mask 255.255.255.0
with default gateway 192.168.80.254. The board can also be assigned to network region 1.
802.1p/Q tagging for an IP board is also enabled or disabled on this form. A number (including 0) in the
VLAN field indicates the VID, and it means that tagging is enabled on the board with that VID.
Although most implementations where tagging is enabled should use VID 0, other VIDs are permitted as
well. The letter ‘n’ in this column means that tagging is disabled on the board, and a blank means that
tagging is not supported on the board. To properly enable L2 tagging on the C-LAN and MedPro/MR320
boards, follow the instructions in section 2.3 under the heading “Rules for 802.1p/Q Tagging.”
The speed and duplex settings for an IP board are configured on this form under the Ethernet Options
heading.
The TN2602AP MR320 board has a VOIP Channels parameter to indicate how many channels are active
on the board. While this parameter is configurable, it is restricted by licensing. The initial licensing
options are to purchase a number of boards with 80 channels each, and a number of boards with 320
channels each.
The C-LAN board has the parameter Number of CLAN Sockets Before Warning. This is related to the
information in section 3.4, heading “C-LAN Capacity and Recommendations.” This parameter only
dictates when a warning is triggered and does not affect the total number of TCP sockets supported by the
C-LAN. Although the recommended number of sockets on a C-LAN may be less than 400, it is advisable
in many cases to wait until 400 (default value) to trigger an alarm.
ip-codec-set
Options are change, display, and list. This form is used to define the codec sets that are referenced by
other IP telephony forms. Up to 7 codec sets may be defined with 5 codecs, listed in order of preference,
in each set. G.711 (uncompressed) and G.729 (compressed) are the recommended codecs for LAN and
WAN, respectively. No silence suppression and 20-ms voice packets are also recommended.
Note about silence suppression: Although silence suppression conserves bandwidth by not transmitting
audio packets during periods of silence, its use typically results in audio clipping, which most users
consider unacceptable. The G.729B codec may be a better alternative to silence suppression. Rather than
not transmitting during silence, this codec transmits silence in a condensed format that requires less
bandwidth. The audio quality of G.729B is still noticeably inferior to G.729.
Larger packet size = less bandwidth Note about voice packet size: Audio is encoded in
Smaller packet size = more bandwidth increments called frames, with the typical frame size
being 10ms. The packet size, or number of frames per
Larger packet size → low loss, high jitter packet, is a measure of how much audio is sent in each
network IP packet. Experience has shown that a 20-ms packet
Smaller packet size → high loss, low jitter is a good compromise between audio quality and
network bandwidth consumption. Reducing to 10ms doubles
the number of packets put onto the network, but only
20-ms packet size recommended 10ms of audio can be lost when a packet fails to reach
its destination or arrives out of order. Going beyond
20ms reduces the number of packets put onto the network, but there is greater potential for poor audio
quality when there is high packet loss.
Larger packets work better in low loss, high jitter networks. Smaller packets work better in high loss, low
jitter networks. 20-ms packets are a good compromise.
The Media Encryption portion of this form is an ordered list of preferred media encryption options. For
example, an ordered list of AES, AEA, and none means that AES encryption is preferred first, then AEA
encryption if AES is not possible, then no encryption if neither AES nor AEA is possible. This list may
contain one or more items.
Allow Direct-IP Multimedia has to do with video over IP, which is beyond the scope of this document.
For information on the remaining FAX, Modem, TDD/TTY, and Clear-channel parameters, see the
product documentation “Administration for Network Connectivity for Avaya Communication Manager”
(555-233-504), chapter 3, heading “Administering FAX, modem, TTY, and H.323 clear channel calls
over IP trunks.” See also the document “Avaya FoIP, MoIP, & TTYoIP” at www.avaya.com.
The Location parameter is used to assign IP stations in this network region to a specific geographic
location identifier.
The Authoritative Domain applies to Session Initiation Protocol (SIP) applications, which are not covered
in this document.
The Codec Set refers to one of the seven codec sets defined using the ip-codec-set form, and specifies
which codec(s) are used by the endpoints in this network region.
The UDP Port Min/Max is the range used for RTP audio by the MedPro and MR320 boards and VoIP
media modules in this network region. Use the following points to configure a more narrow UDP port
range (to set up security filters, for example).
- 2048 is the beginning of the range by default, but this can be changed to a higher starting point.
- The MedPro supports 64 uncompressed audio streams (G.711 codec) or 32 compressed audio streams
(G.729 codec) or any combination using the following formula: [uncompressed streams +
2(compressed streams)] = 64. The MR320 supports up to 320 audio streams, depending on licensing
and configuration.
- Per the RTP standard, each audio stream requires an even-numbered UDP port for the RTP audio, and
the subsequent odd-numbered UDP port for the RTCP control exchange.
- Therefore, to support X audio streams the UDP port range must contain 2X consecutive ports,
beginning with an even port and ending with an odd port. Since the absolute maximum value for X is
320 (MR320 board), the largest required UDP port range is 640.
The DiffServ (DSCP) and 802.1p/Q parameters are the L3 and L2 priority values for call signaling from
C-LANs in this network region, and audio from MedPros/MR320s in this network region. The L2 values
are only applied to boards that have L2 tagging enabled via the ip-interface form. The reason for the two
forms is that L2 tagging and VID can vary per board across a network region, but the priority values are
typically uniform throughout the region.
Ideally two different sets of L2/L3 values should be specified for signaling and audio. However, for
practical purposes in many applications it is common to use the same set of values for both signaling and
audio. Appendix F gives examples of how the L3 values are used in conjunction with QoS on routers.
L2 and L3 prioritization on the C-LAN requires the TN799DP board with firmware v5 or later.
Direct IP-IP Audio (shuffling) and IP Audio Hairpinning within a network region and across different
network regions are enabled and disabled on this form. Direct IP-IP audio permits calls between IP
endpoints to “shuffle” directly to each other, instead of speaking through the MedPro/MR320 board or
VoIP module. If a feature that requires the media gateway, such as conferencing, is activated during the
call, the endpoints shuffle back to the MedPro/MR320 board or VoIP module. If then the conference
ends and only two parties remain, the IP stations shuffle back to one another.
Direct IP-IP Audio and IP Audio Hairpinning are generally enabled, unless there is an Avaya R300 or
MultiVOIP gateway in this network region, in which case hairpinning should be disabled. Also, for direct
IP-IP audio to function across different network regions, an inter-region codec set must be specified and
the regions must be connected via the inter-region connectivity matrix beginning on page 3 of this form.
There are network address translation (NAT) options for direct IP-IP audio. Since Avaya Communication
Manager 1.3, Avaya has permitted shuffling between endpoints that are separated by NAT. NAT has
been a hurdle for VoIP due to the fact that the address in the IP header is translated, but embedded IP
addresses in the H.323 messages are not translated. This hurdle has been overcome to some extent with
the “NAT shuffling” feature in Communication Manager, without the need for H.323-aware NAT
devices. See “NAT Tutorial and Avaya Communication Manager 1.3 NAT Shuffling Feature” at
www.avaya.com.
Note: In addition to the ip-network-region form, shuffling and hairpinning must be enabled on two other
forms: the system-parameters features form, page 16; and the station form, page 2, for each station.
The RTCP monitoring feature is used with the Avaya VoIP Monitoring Manager (VMM). Enabling this
feature causes the audio endpoints in this region to send periodic RTCP reports to VMM. VMM uses
these reports to keep a history of audio quality for all reporting endpoints. The default server parameters
are configured on the system-parameters ip-options form. If the default settings are not desired in any
given network region, specific settings can be applied on a per region basis.
The RSVP feature requires careful integration with the IP network and must not be enabled without the
supporting IP network configurations. These configurations can be cumbersome and require a significant
amount of network overhead. A better call admission control (CAC) mechanism is native to
Communication Manager as of 2.0 and is explained in detail in the “Avaya Communication Manager
Network Region Configuration Guide” at www.avaya.com.
The H.323 Link Bounce Recovery parameters, the LSP list on page 2 of this form, and the inter-region
connectivity matrix beginning on page 3 of this form are covered in detail in a separate document. See
the “Avaya Communication Manager Network Region Configuration Guide” at www.avaya.com.
Inter-Gateway Alternate Routing (IGAR) on page 2 of this form is a new feature for Communication
Manager 3.0. This feature is covered in detail in the “Avaya Communication Manager Network Region
Configuration Guide” at www.avaya.com. Related to IGAR is a new parameter on the cabinet form to
assign the cabinet to a network region. The assignment of a cabinet to a network region, which is a
concept new to Communication Manager 3.0, applies primarily to IGAR. It has no relation to IP boards
in that cabinet, and it does not assign traditional resources attached to that cabinet, such as non-IP stations
and trunks, to a network region.
ip-network-map
Options are change and display. This form is used to assign stations to Communication Manager
network regions by IP address range or subnet. If a station’s IP address does not fall into any of the
ranges configured on this form, the station is assigned to the same network region as the gatekeeper it
registers with. Whether by assignment on this form or by inheritance, it is very important to assign IP
stations to the proper network region. To understand how these methods of network region assignment
The VLAN column is used to send a VID to IP phones. This field should only be used if DHCP option
176 is not available. If such is the case, then two rows are required on this form: one row for the data
VLAN through which the phone passes, and another row for the voice VLAN on which the phone finally
resides, with both rows containing the voice VID. The resulting functionality would be as follows.
- IP phone boots and obtains address on data VLAN.
- IP phone registers with Communication Manager from data VLAN.
- ip-network-map shows phone assigned to a specific network region on a specific voice VLAN.
- Communication Manager directs phone to that voice VLAN.
- IP phone releases data VLAN address and obtains address on voice VLAN.
- IP phone registers with Communication Manager from voice VLAN.
- ip-network-map shows phone assigned to a specific network region on a specific voice VLAN.
- Communication Manager directs phone to that voice VLAN, but phone is already on it.
- Using this method, the phone applies the L2 priority values for audio and signaling as administered
on the ip-network-region form for the phone’s region. Using the recommended DHCP option 176
method, the phone applies the L2 priority values received from DHCP.
The Emergency Location Extension is part of the E911 features of Communication Manager and is not
within the scope of this document.
station
Options are add, change, display, and list. This form is used to define stations. To specify an IP station
the Type must be an IP model. The Port is automatically set to X for an IP phone when the station is first
added. This is changed to S##### – an automatically assigned internal port number – when the phone
registers with the call server. The IP Softphone inquiry is regarding whether or not a softphone is
permitted to take over the extension. This field applies to non-IP stations as well, as an IP softphone can
take over an analog or DCP extension and emulate that set type. Survivable GK Node Name provides an
option for the station to fail over to an Avaya G150 Media Gateway or a MultiTech MultiVOIP gateway
when no other gatekeeper is available. Direct IP-IP Audio and IP Audio Hairpinning for the individual
station is configured on page 2 of this form.
On the trunk-group form, the Group Type should be isdn, the Carrier Medium should be IP, and each
member’s Port designation (beginning on page 3 of the form) should also be IP. Once the members are
used for active calls the call server automatically changes the port designations to T#####, which are
internal port numbers. The number of members determines the number of simultaneous calls.
The LRQ Required parameter allows IP trunk availability to be determined on a per call basis. When this
option is enabled a RAS-Location Request (LRQ) message is sent to the far-end gatekeeper prior to each
call over the IP trunk. The far-end gatekeeper responds with a RAS-Location Confirm (LCF) message
and the call proceeds. The absence of an LCF from the far-end gatekeeper indicates that the call cannot
proceed. If this occurs and the near-end gatekeeper is configured with the necessary route pattern, the
next preferred trunk in the route pattern is used for that call as follows.
- Send LRQ.
- Wait 2sec for LCF (1sec as of Communication Manager 3.0).
- Send LRQ.
- Wait 2sec for LCF (1sec as of Communication Manager 3.0).
- Go to next preferred trunk in route pattern (4sec total per call for Communication Manager pre-3.0;
2sec total per call as of 3.0).
The LRQ feature affects individual calls, whereas the IP trunk bypass feature affects entire IP trunks. The
IP trunk bypass feature takes some time to detect a problem in the IP network and put the signaling-group
into bypass state. When this happens, with the appropriate route pattern in place, it results in all calls
being routed onto the next preferred trunk. The LRQ feature speeds up per call re-routes until IP trunk
bypass is established, so the two features can work in conjunction.
For information about IP trunking with the Cisco Call Manager, see “Avaya S8300 Media Server and
Avaya S8700 Media Server Networked with Cisco Call Manager using H.323 Signaling and IP Trunk
Groups” at www.avaya.com.
media-gateway
Options are add, change, display, and list. This form is used to administer a G700/G350/G250 media
gateway. Number is simply a numeric index. Type is the media gateway model (ie, G700, G350, G250,
G250-BRI). Name is a text descriptor. Serial No is the gateway’s serial number, which is displayed by
typing show system at the MGP CLI. A gateway must be administered on the call server before it can
register to that server, and the serial number is what uniquely identifies a valid gateway.
Network Region is used for IGAR purposes, similar to assigning port networks to a network region on the
cabinet form. But unlike the cabinet form, the network region designation on the media-gateway form
also assigns the gateway VoIP resources to a particular Communication Manager network region. This is
equivalent to assigning a MedPro/MR320 board to a network region on the ip-interfaces form. Recovery
Rule determines automatic recovery back to the primary server while the media gateway is registered to
an LSP. The default is no automatic recovery (‘none’), or a number can be placed here to apply a
recovery rule, per the system-parameters mg-recovery-rule form, as explained in the following section.
Encrypt Link refers to the H.248 signaling link between the gateway and the call server.
Location serves the same function as the identical field on the cabinet form; it is used for call routing
purposes (see the “Avaya Communication Manager Network Region Configuration Guide” at
www.avaya.com). Site Data can be used to note the gateway’s address (ie, if it is located at a remote
branch office). For G250 models Max Survivable IP Ext refers to how many IP stations are permitted to
fail over to the gateway when connectivity to the primary call server is lost. This is part of the SLS
feature, new to Communication Manager 3.0 and the G250. The remaining information is automatically
populated when the gateway registers with the call server.
system-parameters mg-recovery-rule
Options are change and display. When a media gateway loses connectivity to the primary call server, it
can fail over to an LSP. This form, new to Communication Manager 3.0, administers rules that determine
when a media gateway automatically recovers back to the primary server. The Number is simply a
numeric index. Rule Name is a text descriptor. Migrate H.248 MG to primary and Minimum time of
network stability are the two conditions that must be met before the primary Communication Manager
server accepts a media gateway recovery registration.
First the minimum network stability time condition must be met. Then the recovery can happen…
- Immediately.
- When there are no active calls on the media gateway.
- During a specified time window.
- Either when there are no active calls, or during a specified time window.
A blank Migrate H.248 MG to primary field indicates that the rule is disabled. The failover to an LSP,
and recovery back to the primary server, are covered in detail in the “Avaya Communication Manager
Network Region Configuration Guide” at www.avaya.com.
IP Media Packet Performance Thresholds: These parameters, detailed in Appendix G, are for the IP trunk
bypass feature described in the section covering the signaling-group form.
RTCP Monitor Server: These are the VoIP Monitoring Manager server settings applied to all network
regions, unless specified otherwise in the ip-network-region form.
Automatic Trace Route on Link Failure: This feature relates to the following links.
- Port network control link between S87xx/S8500 server and IPSI board.
- H.248 media gateway control link between CLAN/S8300 and media gateway.
- IP trunk between near-end system and far-end system.
When this feature is enabled, and Communication Manager detects a failure on one of these links,
Communication Manager launches a trace route from the source of the link to the destination of the link.
A failed trace route might indicate that the link failure was associated with a network fault, whereas a
successful trace route might indicate otherwise. This feature should be disabled if ICMP is blocked on
the network, so as not to give false indications. The results of this trace route are logged on the call
server, with an IPEVT tag (one of many events with that tag).
H.248 Media Gateway and H.323 IP Endpoint: See the “Avaya Communication Manager Network
Region Configuration Guide” at www.avaya.com for information on most of the parameters under these
headings. Only the Periodic Registration Timer is covered here. This timer determines the frequency at
which a forcefully unregistered IP phone attempts to re-register. The primary application is for a desktop
IP telephone that is forcefully unregistered because a user from home takes over the extension with a
softphone. At some point the user logs off the softphone, leaving the extension free for the IP phone to
reacquire. However, the IP telephone doesn’t know when the softphone logs off, so the IP phone simply
attempts to register periodically, and succeeds only after the softphone logs off. This timer determines
that frequency, and it requires IP telephone 2.1 or later.
Music on Hold: This feature applies to media gateways and to port networks in IP-Connect systems with
no traditional PNC (Center Stage or ATM). When music must be delivered via IP between media
gateways and port networks, the music should be transported via the G.711 codec for quality reasons. If
network region assignments are such that there is always a G.711 path between media gateways and port
networks, this feature is not necessary. In some configurations there may not be a G.711 path, and in
such cases setting this parameter to ‘y’ forces the use of G.711 for music transport.
IP DTMF Transmission Mode: The intra-system parameter determines how DTMF tones are passed
within a system between media gateways and IP-connected port networks with no traditional PNC
(Center Stage or ATM). The inter-system parameter, configured on the signaling-group form,
determines how DTMF tones are passed between systems across IP trunks. Note that both ends of the IP
trunk must be configured the same.
The primary issue driving these parameters is the fact that DTMF tones are not accurately reproduced
using compressed codecs. This is particularly an issue for systems that rely on DTMF tones for
functionality. The options operate as follows.
- in-band: If the configured codec is G.711 or G.729, the tones are passed in-band. Otherwise, the
tones are passed out-of-band via call signaling. G.711 accurately passes DTMF tones, while G.729
can pass the tones but is susceptible to error.
The last two options require the MedPro/MR320 board and VoIP media module to detect the tones and
remove them from the outgoing audio stream. Then a message is sent to the call server for each digit to
be sent out-of-band, or a separate RTP packet with the specified payload format is created for each digit.
status station <ext> Gives static view of a station’s status (multiple pages).
list trace station <ext> Gives real-time view of a station’s activities – for tracing calls.
list trace ras ip-stations <ext> Traces a station’s registration events (GRQ, GCF, RRQ, RCF).
status ip-board <slot #> Gives Ethernet interface in/out statistics for an IP board.
status clan-port <slot #17> (ie, 01a0517) Gives C-LAN board statistics (multiple pages).
status clan-usage Gives C-LAN socket usage.
status media-processor all | board <slot #> Gives MedPro/MR320 status for all boards or individual board.
This section covers some general information regarding various Avaya 4600 Series IP Telephone models.
More specific information is available in the “4600 Series IP Telephone LAN Administrator's Guide” and
other IP telephone guides at support.avaya.com. The current GA firmware releases can be obtained at the
same site. Be sure to read the “readme” files that accompany each firmware package.
Note: For simplicity in many IP telephone applications a C-LAN is often called a gatekeeper, although
the call server is the gatekeeper and the C-LAN is only a front end to the gatekeeper.
4.1 Basics
4606/12/24 Speed/Duplex
The integrated hub in the 4606/12/24 models operates at 10M or 100M half duplex. There are generally
no speed/duplex issues with these models. When connected to an Ethernet switch port configured to
auto-negotiate, the Ethernet switch port stabilizes at 100/half. The exception to this is if a PC is attached
to the phone that is capable of only 10M, in which case all three devices stabilize at 10/half. If no PC is
to be attached to the phone, or if the attached PC will always be capable of 100M operation, then it is
good practice to lock down the Ethernet switch port to 100/half. If a PC may be attached to the phone,
and there is a chance that it may have a 10M NIC, leave the Ethernet switch port in auto-negotiate mode.
Note about Single-Speed Bus: Dual-speed hubs and switches must inherently buffer and discard traffic
because of the inconsistent flows (one port receives at 100M but the other can only send at 10M). The
4606/12/24 models are designed with a single-speed bus in the hub and do not perform these functions.
Instead, these functions are transferred to the enterprise Ethernet switch, where they really belong.
Although the IP telephone can accommodate a second user device (the phone itself being the first), its
primary function is not that of an enterprise network device.
The 30A base switch is a 3-port switch integrated into the base stand of 4612 and 4624 sets. The pigtail
cable attaches to the IP phone’s uplink port. The other two ports are an uplink port to connect to the
enterprise Ethernet switch and a user port to connect to a PC, just like the IP telephone. Both ports are
wired such that they require straight Ethernet cables, just like the IP telephone. Each port supports 10/100
operation at full or half duplex. The ports are in auto-negotiate mode and cannot be configured.
Therefore, the attached devices must also be in auto-negotiate mode, or they must be fixed at 100/half or
10/half. Experience has shown that the 30A functions adequately with the attached devices in auto-
negotiate mode. Because the 30A is not an enterprise-class switch, it is best to have the speed and duplex
on both ports be the same. Otherwise the 30A will be required to buffer and discard frames, which it
cannot do as well as an enterprise Ethernet switch.
The 30A also does something interesting in terms of 802.1Q tagging. It strips the tag from the IP
telephone toward the PC. That is, tagged traffic from the phone is sent to the Ethernet switch (uplink
port) with the tag, but to the attached PC (user port) without the tag. This allows the attached PC to
communicate with the IP telephone when they are on the same VLAN and the phone is tagging. This is
not the case when the PC is connected directly into the phone’s hub port, because the hub port does not
strip the tag, and most PCs do not interpret the tag correctly.
The built-in switch of the 4620 and 4610 operates much like the 30A base switch in terms of
speed/duplex, QoS, and treatment of tagged traffic from the phone. Additionally, the built-in switch
permits speed and duplex configuration if necessary. Follow the guidelines in section 2.1, heading
“Speed/Duplex” when configuring the speed and duplex on these phones and the Ethernet switch ports to
which they are connected.
Other than the hub/switch difference between the legacy and current models, they are similar in terms of
network-related implementation. They are different in terms of feature-related implementation, and the
additional features and functionality of the current models are covered in IP telephone specific
documentation found at support.avaya.com.
The DHCP specification has what are called options, numbered from 0 through 255. Each option is
associated with a specific bit of information to be sent by the DHCP server to the DHCP client. For
example, option 1 is the subnet mask option and is used to send the subnet mask to the client. Option 3 is
the router option and is used to send the default gateway address and other gateway addresses to the
client. Some options are defined – such as options 1 and 3 – and others are not. The defined options are
found in RFC 2132.
Options 128 through 254 are site-specific options. They are standard options that are not defined, and
vendors may use these options and define them to be whatever is necessary for a specific application.
Avaya IP telephones use site-specific option 176 as one of the methods to receive certain parameters from
the DHCP server.
Parameter Value
MCIPADD Address(es) of gatekeeper(s) – at least one required
MCPORT The UDP port used for registration – 1719 default
TFTPSRVR Address(es) of TFTP server(s) – at least one required
L2QVLAN 802.1Q VLAN ID – 0 default
L2QAUD L2 audio priority value.
L2QSIG L2 signaling priority value.
VLANTEST The number of seconds a phone will attempt to return to the
previously known voice VLAN
Table 8: DHCP option 176 parameters and values
The typical option 176 string for a single-VLAN environment looks like this.
MCIPADD=addr1,addr2,addr3, … ,MCPORT=1719,TFTPSRVR=addr
At least one gatekeeper (C-LAN or S8300) address must be present after MCIPADD to point the phones
to a call server. MCPORT specifies which UDP port to use for RAS registration. IP telephone firmware
1.6.1 and later already have 1719 as the default port, but it is prudent to include it. A TFTP server
address is necessary so that phones know where to go to download the necessary script files and binary
codes (see “Boot-up Sequence” heading below). L2QVLAN and VLANTEST would be included if
802.1Q tagging were required, such as in a dual-VLAN environment (see section 4.2). Other parameters
may be added, such as L2QAUD and L2QSIG, which are used to specify the L2 priority values for audio
and signaling. If these values are not specified in option 176, the default values (6/6) are used.
Note: The L3 priority values (DSCP) are received from the call server, as configured on the SAT ip-
network-region form. The reason L3 values are received from the call server and L2 values are not is
because an IP phone accepts all L2 values from one source. The preferred and recommended method is
via DHCP option 176. An alternative method is described in section 3.5, heading “ip-network-map,”
which utilizes the L2 values administered on the SAT ip-network-region form.
An administrator must create option 176 on the DHCP server and administer a properly formatted string
with the appropriate values. Option 176 could be applied globally or on a per scope basis. The
recommendation is to configure option 176 on a per scope basis, because the values themselves or the
order of the values could change on a per scope basis. As part of the DHCP process at boot-up, the IP
telephone requests option 176 from the DHCP server.
To specify TLS or HTTP script/firmware downloads, in option 176 of the DHCP scope apply the
TLSSRVR (for TLS) or HTTPSRVR (for HTTP) parameter in lieu of TFTPSRVR. If TLSSRVR,
HTTPSRVR, and TFTPSRVR are all set, the phone will attempt to download firmware using TLS first on
TCP port 411, then HTTP on TCP port 81, then HTTP on TCP port 80, then TFTP on UDP port 69.
Note: Avaya IP telephones only establish encrypted TLS connections with servers using an Avaya-signed
digital certificate (ie, an Avaya S8300 or S8500 Media Server).
Boot-up Sequence
The following are key boot-up events, listed in order, which may help to verify proper operation of the IP
phone. This list may not be comprehensive, as only key events are listed. The packets described here can
be captured using a protocol analyzer, and one with H.323 capability is required to properly decode the
H.225 RAS messages. On 4606/12/24 models the analyzer can be attached to the phone’s user port. But
because the 4620 and 4610 have a built-in switch instead of a hub, the analyzer must be attached to a
mirrored Ethernet switch port, or to a tap or hub in line between the phone and the Ethernet switch.
- Initial startup – At power-up or manual reset, the phone goes through a short initial startup procedure.
The display shows Restarting… (if the phone was intentionally restarted w/ Hold RESET#), and
then Loading… and Starting…
- DHCP – The phone queries the DHCP server for an IP address and other needed information. The
following packets are exchanged: DHCP Discover from phone to broadcast; DHCP Offer from server
to broadcast, or relay agent to phone; DHCP Request from phone to broadcast; and DHCP ACK from
server to broadcast, or relay agent to phone. Note that this step is bypassed if the phone is manually
configured with all the necessary information.
- Request file “46xxupgrade.scr” and others from TLS/HTTP/TFTP server – This is a text script file
that tells the phone which boot code and application code are needed. If the phone does not have the
current codes, it requests them from the file server. A brand new phone makes all three requests, as
phones typically come from the factory with outdated code. In addition, the “46xxupgrade.scr” script
may instruct the phone to download the “46xxsettings.scr (or .txt)” file, which is an optional method
of sending configurations to the phone. Note that there is a loading period after each .bin code is
received for the first time. Note also that the file names are case sensitive on some servers
(Unix/Linux) and not on others (Microsoft).
- Ext and Password prompts – The phone prompts for the extension and password if there are no
previously stored values.
- Registration with gatekeeper – The phone registers with a gatekeeper (C-LAN or S8300) after the
extension and password are entered. This registration happens very quickly and does not show up on
the display. However, the following packets are exchanged: RAS-Gatekeeper Request (GRQ) from
phone to gatekeeper; RAS-Gatekeeper Confirm (GCF) from gatekeeper to phone; RAS-Registration
Request (RRQ) from phone to gatekeeper (not necessarily the same one GRQ was sent to); RAS-
Registration Confirm (RCF) from gatekeeper to phone.
- H.225 call signaling connection – The phone opens a TCP session with the gatekeeper and sends an
H.225 Setup message, which is answered with H.225 Call Proceeding and H.225 Connect messages
from the gatekeeper. This call signaling session remains up throughout the registration. During idle
periods the phone maintains the session by sending TCP keepalives.
Call Sequence
It is not feasible to give a standard packet-by-packet call sequence, because of the many possible
variations on any given call. Instead, a higher level description of the process is offered here. Depending
on which features are enabled and executed during a call the packet-by-packet sequence may vary, but the
fundamental functions described here apply overall. All call signaling functions go through the
gatekeeper, either via the C-LAN or natively (S8300), and the gatekeeper dictates what the IP stations do
during a call.
- Calling phone contacts gatekeeper on already established call signaling session (TCP 1720 gatekeeper
port, variable phone port).
- There are some call signaling exchanges on this TCP session.
- Calling phone establishes an audio stream with an audio resource (MedPro/MR320 board or VoIP
module), as directed by the gatekeeper.
- Gatekeeper contacts called phone on already established call signaling session (TCP 1720 gatekeeper
port, variable phone port).
- There are some call signaling exchanges on this TCP session.
- Called phone also establishes an audio stream with an audio resource, as directed by the gatekeeper,
but this stream is one-way until the call completes.
- Called phone answers, resulting in more call signaling activity, and the call completes. The call could
remain in this state, but…
- In most cases, unless configured otherwise, the gatekeeper contacts both phones and instructs them to
direct their audio streams to each other.
- Phones direct audio streams to each other, as instructed by the gatekeeper.
- One of the phones hangs up, resulting in more call signaling activity.
- Gatekeeper contacts both phones, signals that the call has ended, and instructs them to tear down
audio streams.
- Phones tear down audio streams.
Keepalive Mechanisms
There are two types of keepalive mechanisms: RAS and TCP.
- RAS keepalive – The IP telephone sends RAS keepalive messages to the gatekeeper at a time-to-live
(TTL) interval specified by the gatekeeper. On a protocol analyzer a RAS keepalive message shows
up as a RAS-Registration Request (RRQ) with the keepalive bit set in the RAS decode. Each request
message is acknowledged by the gatekeeper with a RAS-Registration Confirm (RCF). This exchange
takes place over the RAS socket, which has UDP port 1719 on the gatekeeper side.
- TCP keepalive – The IP telephone sends TCP keepalive messages to the gatekeeper at a regular
interval determined by the phone, or as administered on the ip-network-region form. The keepalive
is an empty TCP datagram with a sequence number that is 1 to 5 less than the sequence number of the
previous real TCP message or ACK sent by the phone. The gatekeeper acknowledges each keepalive
from the phone with a similar empty TCP datagram. This exchange takes place over the call
signaling socket, which has TCP port 1720 on the gatekeeper side.
- Regular and retry intervals – Each keepalive mechanism has a regular interval as described above.
If a regular interval keepalive is not acknowledged, more keepalives are sent at a faster retry
Independent of the mechanism (RAS or TCP), the keepalive flow follows this pattern.
regular interval regular interval retry int retry int retry int retry int retry int retry int
discovering
failure
The discovering at the end of the flow means that the phone has effectively unregistered and is searching
for another gatekeeper. Effectively unregistered means that the phone has not sent an explicit RAS-
Unregistration Request (URQ) message, but it considers itself unregistered from that gatekeeper and is
moving on to the next. Even if the phone did send a URQ, chances are the gatekeeper would not receive
it because the failure condition could still exist.
The final retry interval prior to discovering would appear to give extra time for the failure to recover.
And indeed if the phone did receive a KA acknowledgment within that final retry interval it would stay
registered to the same gatekeeper. However, the reality is that if the phone doesn’t receive an
acknowledgment within a second or two after the final retry KA, it won’t receive one. Therefore, the
final retry interval really does not factor into the time to unregister. Time to unregister answers the
question, “How long must the failure (ie, network outage) last before the IP telephone unregisters?” If the
failure recovers just before the final retry KA is sent, the phone remains registered to the same
gatekeeper. If the failure recovers a couple seconds after the final retry KA is sent, the phone most likely
unregisters and moves on to the next gatekeeper after the final retry interval.
On the back of the phone, the port with the icon that looks like a terminal is the user port. (The port with
the icon that looks like a network jack is the uplink port, which connects to the Ethernet switch.) Use
discretion when connecting a PC to the phone, and remember that its primary function is not that of an
enterprise network device. For example, do not connect an enterprise server to the phone. Such high-
traffic servers require their own separate connections to the enterprise Ethernet switch. Also, do not
connect a PC to the phone with a 10M uplink to the network. The phone itself operates well at 10M, but
with a PC attached the two should operate at 100M.
The second scenario is similar to the first, except that traffic from the phone is marked with L2 and/or L3
priority while remaining on the port/native VLAN. See the instructions in section 2.3 under the heading
“Rules for 802.1p/Q Tagging.” The phone must be configured to apply the appropriate L2 and/or L3
priority values. The Hold ADDR# menu is used to manually enable or disable 802.1Q tagging and to set
the VLAN ID. The other parameters are configured via the Hold QOS# menu. The manual method is
covered below, and an automated method is covered in the next paragraph.
- 802.1Q – On/off for 802.1Q tagging. Turn this on if L2 priority tagging is desired; off otherwise.
- VLAN ID – Should be zero (0) for this scenario, per the instructions in section 2.3, heading “Rules
for 802.1p/Q tagging.” The VID has no effect when 802.1Q tagging is disabled.
- VLANTEST – Not relevant when VID is zero. Applies in a dual-VLAN environment when VID is a
non-zero value, as explained in later sections.
- L2 audio – Layer 2 CoS tag for Ethernet frames containing audio packets. The phone either receives
this from DHCP (most common) or from the call server (rare), per the ip-network-region form. This
value could also be set manually on a per phone basis.
- L2 signaling – Layer 2 CoS tag for Ethernet frames containing signaling packets. The phone either
receives this from DHCP (most common) or from the call server (rare), per the ip-network-region
form. This value could also be set manually on a per phone basis.
- L3 audio – Layer 3 DSCP for audio IP packets. The phone automatically receives this value from the
call server, per the ip-network-region form. This value could also be set manually on a per phone
basis.
- L3 signaling – Layer 3 DSCP for signaling IP packets. The phone automatically receives this value
from the call server, per the ip-network-region form. This value could also be set manually on a per
phone basis.
The manual menus are covered here for explanatory purposes. However, a better alternative is to use
DHCP option 176 and the built-in capabilities of the call server and IP telephone to automatically
configure the phones. As stated previously, the call server sends the L3 priority values to the phones
automatically, per the values configured in the ip-network-region form. The 802.1Q on/off instruction,
VLAN ID, and L2 priorities can be configured automatically using DHCP option 176 as described in
section 4.1, heading “DHCP Option 176.” Here is what that string should look like for 1.8 and later
phones (see the appropriate “LAN Administrator’s Guide” for previous phone releases).
MCIPADD=addr1,addr2, … ,MCPORT=1719,TFTPSRVR=addr,L2QVLAN=0,L2QAUD=#,L2QSIG=#
Remember that in order for the CoS markings to have any effect, the corresponding QoS configurations
must be implemented on the necessary network devices. Remember also that improperly enabling L2 and
L3 prioritization may break processes that were working without it. Read section 2.3 of this document for
more information on CoS and QoS.
The Hold ADDR# and Hold QOS# menu options are the same as described in the previous heading,
except that the VID must not be zero. The preferred method of using DHCP option 176 (section 4.1,
heading “DHCP Option 176”) is also the same, except that L2QVLAN has a non-zero value. Finally, in
a dual-VLAN implementation the VLANTEST parameter has great significance, as illustrated below.
The following scenario, with arbitrary voice VLAN ID, details the steps a phone (1.8 and later) would go
through in a typical dual-VLAN implementation. It also illustrates the recommended content of the
option 176 string.
- Phone with no previously stored values boots up and obtains an address on the data VLAN.
- The data VLAN option 176 string directs the phone to go to voice VLAN 25.
MCIPADD=addr1,addr2, … ,MCPORT=1719,TFTPSRVR=addr,L2QVLAN=25,L2QAUD=6,L2QSIG=6,VLANTEST=600
- Phone releases the data VLAN address and obtains an address on the voice VLAN.
- The voice VLAN option 176 string is identical to the data VLAN string but without the L2QVLAN
parameter, because a phone already on the voice VLAN doesn’t need to be directed to go there.
MCIPADD=addr1,addr2, … ,MCPORT=1719,TFTPSRVR=addr,L2QAUD=6,L2QSIG=6,VLANTEST=600
- Phone is operational on the voice VLAN.
- Reboot or power cycle occurs.
- Phone immediately returns to voice VLAN 25 upon recovery, and one of the following occurs.
- Phone obtains an address and option 176 string on the voice VLAN and all is well.
- Phone cannot obtain an address on the voice VLAN, due to network or DHCP problems. In this
case the VLANTEST=600 parameter directs the phone to continue trying for 600sec (finite range
is 1-999). If the phone does not succeed in obtaining an address within 600sec, it marks VLAN
25 as invalid and returns to no tagging (back to the data VLAN).
The idea behind going back to the data VLAN after some time is that the phone may have changed ports
and be on one with a different voice VLAN. In such a case the phone would have to start over and be
directed to the proper voice VLAN. The idea behind marking VLAN 25 as invalid in the previous
scenario is that if the phone hasn’t changed ports, it is preferable to operate on the data VLAN than to be
sent to a bad voice VLAN in a continuous loop. For cases where it is not preferable to operate on the data
VLAN, the option VLANTEST=0 was added as of legacy phone firmware 1.8.2 and current phone
firmware 2.0.1. This instructs the phone to permanently remain on the previously known voice VLAN.
Note: DHCP option 176 is the preferred method for directing IP phones to the voice VLAN. The method
described previously using the VLAN field of the ip-network-map form is an alternative if DHCP option
176 is not available. The two methods should not be used simultaneously.
Remember that in order for the CoS markings to have any effect, the corresponding QoS configurations
must be implemented on the necessary network devices. Remember also that improperly enabling L2 and
L3 prioritization may break processes that were working without it. Read section 2.3 of this document for
more information on CoS and QoS.
An IP telephone can have a list of gatekeepers (C-LANs and/or S8300s) to which it may send the initial
RAS-Gatekeeper Request (GRQ) message. This list is obtained via the DHCP option 176 string, which is
covered briefly in section 4.1 and in detail in the “LAN Administrator’s Guide.”
Within the DHCP option 176 string, the comma-separated IP addresses that follow the MCIPADD
parameter constitute a gatekeeper list, and this list provides redundancy at boot-up. If a given gatekeeper
is unreachable for any reason, the phone attempts other gatekeepers in the gatekeeper list. The following
hypothetical network diagram and the accompanying instructions explain how gatekeeper lists should be
administered on DHCP servers.
DHCP Server w/
scopes for v80-90
Router
Network Region 2
Access Access
Switches
Access Switches
Access
AN
Switch Switch
DHCP Server w/
Network Region 1
scopes for v10-40 g700 with
Core Core s8300 LSP
Switch Switch VoIP mod
phones on phones on
v80 get info v90 get info
Distribution Distribution Distribution Distribution from DHCP from DHCP
C-LAN 1 Switch Switch Switch Switch C-LAN 4 v80 scope v90 scope
C-LAN 2 C-LAN 3
Suppose, for whatever reason, that a large number of IP phones are rebooted at once. Which
gatekeeper(s) will they contact first? The correct answer is that they should contact all the gatekeepers in
a distributed fashion. All the phones should not bombard the same gatekeeper at once with GRQs. There
are various ways to configure the gatekeeper lists, and the following is possibly the simplest.
Note that this principle may also apply to multiple TFTP servers.
Branch Site
The branch site is just slightly different in terms of the DHCP scopes, but very different in terms of the
failure scenario and other factors that affect the branch implementation.
The IP telephones at the branch site could access the same four C-LANs shown above, or there could be a
different set of C-LANs (not shown) for the branch IP phones. In either case the DHCP scopes for v80
and v90 should have rotating lists, as at the main site. However, in addition to the list of C-LAN
addresses, the v80 and v90 scopes should also include the S8300 LSP address at the end of the list. This
is because the LSP can take over as the call server for the branch if the WAN link fails. The LSP only
accepts registrations when it is active, so having the LSP in the list does not result in inadvertent
registrations to the LSP.
Because an extended WAN link failure is possible, the branch site should ideally have its own DHCP
server. It makes sense that if there is a redundant call server at the branch, there should also be a
dedicated DHCP server, because IP telephones require both services. For cost and administrative reasons,
however, many will choose not to install a DHCP server at all branch locations. In such cases it is very
Here are some key points regarding the option 176 gatekeeper list and the RCF Alternate Gatekeeper List.
- IP telephone versions prior to 2.0 use both lists simultaneously. GK addresses received from either
method are merged into one list.
- IP telephone 2.0 and later maintain the two lists separately, with only one list active at any given time.
During boot-up the phone uses the list obtained from option 176. After registration the phone uses
the Alternate Gatekeeper List received in the RCF. When a phone is logged off but not rebooted, it
reverts back to the list obtained from option 176.
- The option 176 GK list is recommended, as opposed to manual entry or a single GK address in option
176, because the RCF list is received after registration. If the phone only knows of one GK at boot-
up and that GK is out of service, the phone cannot register and hence cannot get an RCF.
- The Alternate Gatekeeper List sent in the RCF follows a specific algorithm. When an IP phone
registers and its network region is specified in the ip-network-map form, the call server delivers a
list of all gatekeepers in that region, plus directly connected regions (specified in the ip-network-
region form). If an IP phone’s network region is not administered in the ip-network-map form, it
inherits the region of the gatekeeper that receives the registration, and the call server delivers a list of
all gatekeepers only in that region.
- As of Avaya Communication Manager 1.3, the addresses of the LSPs (administered on the ip-
network-region form) in the same network region as the IP phone are also sent in the RCF. As of
Communication Manager 2.0, in addition to the LSPs, the address of the Survivable GK Node Name
(administered on the station form) is also sent in the RCF.
- The combination of Communication Manager 2.x and IP telephone 2.x facilitates a distinction
between primary and secondary gatekeepers in the Alternate Gatekeeper List. During recovery after
an outage, the primary gatekeepers are attempted first for a period of time called the H.323 Primary
Search Time, specified in the system-parameters ip-options form. After this search time expires,
the secondary gatekeepers – LSPs and the Survivable GK – are also included in the search. For a
more detailed discussion see the H.323 Link Bounce section of the “Avaya Communication Manager
Network Region Configuration Guide” at www.avaya.com.
- 1.8.x shows combined list from - 1.8.x shows combined list from
RCF and option 176, or RCF and option 176, or
combined RCF list and combined RCF list and
manually configured manually configured
gatekeeper. gatekeeper.
MIB Object ID Shows gatekeeper to which phone is Shows gatekeeper to which phone
.1.3.6.1.4.1.6889.2.69.1.1.4 currently registered. was last registered.
(endptMCIPINUSE)
Hold ADDR# keypad menu Shows gatekeeper to which phone is N/A
currently registered.
VLAN Defined
With simple Eth-switches, the entire switch is one L2 broadcast domain that typically contains one IP
subnet (L3 broadcast domain). Think of a single VLAN (on a VLAN-capable Eth-switch) as being equivalent to a
simple Eth-switch. A VLAN is a logical L2 broadcast domain that typically contains one IP subnet. Therefore,
multiple VLANs are logically separated subnets – analogous to multiple switches being physically separated
subnets. A L3 routing process is required to route between VLANs, just as one is required to route between
switches. This routing process can take place on a connected router or a router module within a L2/L3 Eth-switch.
If there is no routing process associated with a VLAN, devices on that VLAN can only communicate with other
devices on the same VLAN.
For a tutorial and more information on VLANs, see “LANs and VLANs: A Simplified Tutorial” at
www.avaya.com.
Port VLAN and native VLAN are synonymous terms. The IEEE standard and most Avaya switches use the
term port VLAN [6 p.11], but Cisco switches use the term native VLAN. Issue the command show trunk on Avaya
P330s and Cisco CatOS switches to see which term is used in the display output.
Every port has a port/native VLAN. Unless otherwise configured, it is VLAN 1 by default. It can be
configured on a per port basis with the following commands.
All clear Ethernet frames (ones with no 802.1Q tag, such as from a PC) are forwarded on the port/native
VLAN. This is true even if the Eth-switch port is configured as an 802.1Q trunk, or otherwise configured for
multiple VLANs (see VLAN binding heading below).
Configuring a Trunk
A trunk port on an Eth-switch is one that is capable of forwarding Ethernet frames on multiple VLANs via
the mechanism of VLAN tagging. IEEE 802.1Q specifies the standard method for VLAN tagging. Cisco also uses
a proprietary method called ISL. Avaya products do not interoperate with ISL.
A trunk link is a connection between two devices across trunk ports. This can be between a router and a
switch, between two switches, or between a switch and an IP phone. Some form of trunking or forwarding multiple
VLANs must be enabled to permit the IP phone and the attached PC to be on separate VLANs. The following
commands enable trunking.
By default only the port/native VLAN is enabled on By default all VLANs (1-1005) are enabled on the
the trunk port. Another set of commands is required trunk port. VLANs can be selectively removed with
to specify other allowed VLANs. the command clear trunk <mod/port> <vid>.
On the Avaya P330, additional VLANs are added to a port using the VLAN binding feature. The port may
be a trunk port (802.1Q tagging enabled) or an access port (no 802.1Q tagging). The port does not need to be a
trunk to forward multiple VLANs, and for one application – connecting to an Avaya IP phone – it must not be a
trunk (ie, do not issue the set trunk command). The following steps enable VLAN binding.
1. Verify that the port is configured with the desired port/native VLAN.
2. Add additional VLANs with one of the following vlan-binding-mode options.
Static option:
set port vlan-binding-mode <mod/port> static Put the port in bind-to-static mode.
set port static-vlan <mod/port> <vid> Statically add another VLAN, in addition to the
port/native VLAN.
----- OR -----
Configured option:
set vlan <id> Add a VLAN to the configured VLAN list. Type show
vlan to see entire list.
set port vlan-binding-mode <mod/port> bind-to- Apply the configured VLANs to the port and permit
configured only those VLANs (bind-to-all permits all VLANs and
not just the configured).
3. If the port is connected to a router or to another switch, trunking must be enabled with the command set trunk
<mod/port> dot1q, which causes all egress frames to be tagged. However, if the port is connected to an Avaya
IP phone with an attached PC, trunking must not be enabled so that none of the egress frames are tagged. This
is necessary because most PCs do not understand tagged frames.
With Avaya switches it is possible to set the L2 priority on the IP phone, even if the phone is not connected
to a trunk or multi-VLAN port. That is, the Avaya switch does not need to be explicitly configured to accept
priority-tagged Ethernet frames on a port with only the port/native VLAN configured. This is useful if the phone
and the attached PC are on the same VLAN (same IP subnet), but the phone traffic requires higher priority. Simply
enable 802.1Q tagging on the IP phone, set the priorities as desired, and set the VID to zero (0). Per the IEEE
standard, a VID of zero assigns the Ethernet frame to the port/native VLAN.
Cisco switches behave differently in this scenario, depending on the hardware platforms and OS versions.
Here are Avaya Labs test results with a sample of hardware platforms and OS versions.
Catalyst 6509 w/ Accepted VID zero for the native VLAN when 802.1Q trunking was
CatOS 6.1(2) enabled on the port. In this case, all but the native VLAN should be cleared
off the trunk.
Would not accept VID zero for the native VLAN. Opened a case with
Cisco TAC, and TAC engineer said it was a hardware problem in the 4000.
Catalyst 4000 w/ Bug ID is CSCdr06231. Workaround is to enable 802.1Q trunking and tag
CatOS 6.3(3) with native VID instead of zero. Again, clear all but the native VLAN off
the trunk.
Note that setting a L2 priority is only useful if QoS is enabled on the Eth-switch. Otherwise, the priority-
tagged frames are treated no differently than clear frames.
Sample Multi-VLAN Scenario for Avaya P330 Code 3.2.8 and Cisco CatOS and IOS
Here is a sample multi-VLAN scenario. Suppose there is a Cisco router connected to a P330 switch that
contains two VLANs, one for the VoIP devices and one for the PCs. To conserve ports and cabling, the PCs are
connected to the phones and the phones are connected to the P330 switch.
C-LAN MedPro
vlan 10 vlan 10
192.168.10.1 192.168.10.2
G650 vlan 10
192.168.10.7
f0/1 1/2 1/3
1/1 Avaya 1/5
Avaya
IP Phone
Cisco Router
Cajun P330
192.168.1.254
1/12
192.168.10.254
vlan 1 vlan 1
192.168.1.100 192.168.1.7
DHCP Server PC
TFTP Server
set port vlan-binding-mode 1/5 static Port in static binding mode by default, but command shown.
set port static-vlan 1/5 10 In addition to v1, v10 statically bound to port, but not a trunk port.
set port spantree disable 1/5 Spanning Tree disabled at the port level.
Port 1/12 for the DHCP/TFTP server already has port/native VLAN 1.
set port spantree disable 1/12 Spanning Tree disabled at the port level.
set port vlan-binding-mode 1/1 bind-to- Port bound to configured VLANs 1 and 10.
configured
set trunk 1/1 dot1q Port connected to Cisco router is an 802.1Q trunk port.
set port spantree disable 1/1 Spanning Tree disabled at the port level.
set port vlan-binding-mode 1/5 bind-to- Bound to configured VLANs but not a trunk port.
configured
set port spantree disable 1/5 Spanning Tree disabled at the port level.
interface FastEthernet0/2
switchport access vlan 10 Port/native VLAN changed to 10 on this port.
spanning-tree portfast Spanning Tree fast start feature.
switchport priority default 6 Port native VLAN L2 (802.1p) priority set to 6.
interface FastEthernet0/3
switchport access vlan 10
spanning-tree portfast
switchport priority default 6
interface FastEthernet0/5
switchport trunk encapsulation dot1q 802.1Q trunk port.
switchport trunk native vlan 1 Since most PCs do not understand the tag, the PC’s VLAN must be
the native VLAN. v1 is already the native, but command shown.
switchport trunk allowed vlan 1,10 VLANs 1 and 10 allowed on trunk.
switchport mode trunk Port is in trunk mode.
spanning-tree portfast Spanning Tree fast start feature.
interface FastEthernet0/5 Simpler configuration on newer IOS switches (ie, 3550, 3560).
switchport mode access Access mode; explicit trunking not required.
switchport access vlan 1 Configure the data VLAN.
switchport voice vlan 10 Configure the voice VLAN.
IP phone configuration
This procedure applies regardless of the Eth-switch used. Initially placing the IP phone on VLAN 10
requires two DHCP scopes – one for VLAN 1 and another for VLAN 10. Both scopes should have identical DHCP
option 176 strings, with one exception. The VLAN 1 scope must have the L2QVLAN parameter, and the VLAN 10
scope should not. The following strings apply to phone firmware 1.8 and beyond.
VLAN 1:
MCIPADD=addr1,addr2, … ,MCPORT=1719,TFTPSRVR=addr,L2QVLAN=10,L2QAUD=6,L2QSIG=6,VLANTEST=0
Run the phone through its normal boot-up sequence. It obtains an IP address on VLAN 1 – the port/native
VLAN. When the phone receives the option 176 string above from the VLAN 1 scope, it releases the VLAN 1
address and enters a second DHCP sequence with tagging enabled to obtain a VLAN 10 address. After the phone is
operational on VLAN 10, on subsequent reboots the phone returns to VLAN 10 directly, without passing through
VLAN 1. In this example the VLANTEST=0 option is invoked to make the phone permanently remain on the
voice VLAN. See section 4.2, heading “IP Phone and Attached PC on Different VLANs” for a full explanation of
how the phone operates between the data and voice VLANs, including the use of the VLANTEST parameter.
The L2QVLAN parameter should not be added to the VLAN 10 DHCP scope. This is so that in the event
a phone is connected to a port that has VLAN 10 as the port/native VLAN, it will not receive instructions from the
PC configuration: The PC can be statically addressed with a VLAN 1 address, or it can receive a VLAN 1 address
via DHCP. No special configurations are required.
Interoperability with auxiliaryvlan and voice vlan was successfully lab tested on the following platforms,
with no known issues to date.
Furthermore, Avaya IP phones have been deployed on a broader range of CatOS and IOS platforms by various
Avaya customers, also with no known issues to date.
Therefore, auxiliaryvlan, voice vlan, and explicit 802.1Q trunking are all viable options when a dual-
VLAN environment is required (see Appendix A). It is left to the user to choose the method, keeping in mind that
auxiliaryvlan and voice vlan are Cisco proprietary mechanisms and are not subject to constraint by a standards body
or by Avaya. 802.1Q trunking is well tested, successfully deployed, and defined by a standards body, but the
configuration is not as clean, and trunking on user ports has other network implications.
For IOS-based Catalyst switches, voice vlan is roughly equivalent to auxiliaryvlan. On older IOS platforms
(ie, 2900XL, 3500XL) there appears to be no configuration or functionality benefit to using voice vlan, as explicit
trunking is still required when voice vlan is enabled on these older platforms. On newer IOS platforms (ie, 3550,
3560), however, voice vlan can be enabled without explicit 802.1Q trunking, so there are benefits to using voice
vlan on these newer platforms.
Note that Avaya IP phones do not interoperate with CDP. Therefore, although auxiliaryvlan and voice vlan
can be used, the mechanism of discovering these VLANs via CDP is not supported. The Avaya IP phone can learn
the auxiliaryvlan/voice vlan designation via DHCP option 176, as explained below and in Appendix A.
How it Works
The remainder of this document focuses on auxiliaryvlan (CatOS), but voice vlan (IOS) operates on the
same principles as auxiliaryvlan.
At the heart of Cisco’s auto-discovery feature are Cisco-proprietary mechanisms. The first proprietary
mechanism is CDP (Cisco Discovery Protocol). This is a layer 2 protocol, which means that it works at the Ethernet
level, without requiring IP addresses. Cisco devices identify themselves to other Cisco devices using CDP packets
that contain device- and port-specific information. (CDP packets can be captured and decoded using protocol
analyzers that support CDP.) With the appropriate devices and OS versions, the CDP packets contain information
specific to VoIP and other real-time applications. [1 p.2-22]
The information passed from the Cisco phone to the Catalyst is not of concern. The phone communicates
its specific power requirements to the Catalyst, and the phone can also trigger the Catalyst to send its CDP packet
immediately instead of waiting for the transmit period (60 seconds by default) to recycle. [1 p.2-23]
The auxiliaryvlan is a modified method of implementing 802.1Q trunking, and it may be nothing more than
this. Although testing to date has been positive, Avaya does not know what other mechanisms are or will be
incorporated with this feature, or if they could have any adverse effects on Avaya IP phones. Assuming that an
auxiliaryvlan-enabled port is truly a standard 802.1Q trunk port, the following steps allow Avaya IP phones to work
on Cisco’s auxiliaryvlan.
Application Perspective
Here is the anatomy of a 20-ms G.729 audio packet, which is recommended for use across limited
bandwidth WAN links. Notice that two-thirds of the packet is consumed by overhead (IP, UDP, and RTP), and only
one-third is used by the actual audio.
It is important to understand that all 20-ms G.729 audio packets, regardless of the vendor, are constructed like this.
Not only is the structure of the packet the same, but the method of encoding and decoding the audio itself is also the
same. This sameness is what allows an Avaya IP phone to communicate directly with a Cisco IP phone, or any
other IP phone, when using matching codecs. The packets from the application perspective are identical.
Network Perspective
RTP header compression is a mechanism employed by routers to reduce the 40 bytes of protocol overhead
to approximately 2 to 4 bytes [7 p.1] [2 p.5-14]. Cisco routers employ this mechanism, as does the Avaya
X330WAN router, which is a module for the P330 chassis. RTP header compression can drastically reduce the
VoIP bandwidth consumption on a WAN link when using 20-ms G.729 audio. When the combined 40-byte header
is reduced to 4 bytes, the total IP packet size is reduced by 60% (from 60 bytes to 24 bytes). This equates to
reducing the total VoIP WAN bandwidth consumption by roughly half, and it applies to all 20-ms G.729 audio
packets, regardless of the vendor.
Customers who deploy routers capable of this feature may be able to benefit from it. However, Cisco
recommends caution in using RTP header compression because it can significantly tax the processor if the
compression is done in software. Depending on the processor load before compression, enabling RTP header
compression could significantly slow down or crash the router. For best results, use a hardware/IOS/interface
module combination that permits the compression to be done in hardware [3 QC-333] [5 “RTP Header
Compression and QoS”].
RTP header compression has to function with exactness or it will disrupt audio. If for any reason the
compression at one end of the WAN link and decompression at the other end do not function properly, the result
could be intermittent loss of audio or one-way audio. This has been very difficult to quantify, but there is some
anecdotal evidence. One production site in particular experienced intermittent one-way audio whose cause was very
difficult to troubleshoot and isolate. When RTP header compression was disabled, simply for experimentation
purposes, the audio problems went away.
The Test
This section details the results of a simple RTP header compression test conducted in a lab environment.
Although this test was conducted using Cisco routers, the expected behavior is the same for any router that performs
this function as specified in RFC 2508 [7]. This test was performed in the following lab configuration.
Parascope
WAN probe
- NetIQ Chariot v4.0 was used to simulate VoIP calls between the two endpoints. Chariot v4.0 accurately
simulates the characteristics of various codecs and uses a 40-byte IP/UDP/RTP header.
- Sniffer Pro v3.50.02 was used to capture the sent and received packets.
- The Cisco 3600 had IOS v12.1(2)T and the Cisco 1600 had IOS v12.0(12).
- The Fredericks Engineering Parascope WAN probe was tapped into the V.35 serial link to take bandwidth
measurements.
- This test was performed using PPP encapsulation on the WAN link.
A single call was placed between the Chariot endpoints using the two most common codecs, sending 20-ms
voice packets. Below are the results with and without RTP header compression. Note that these are rough
measurements.
For each codec there was an attempt to verify that the audio packets were received in tact. This was done
by spot-checking the audio packets before and after compression, using two Sniffer protocol analyzers. With G.729
the RTP header and payload were identical before and after compression. With G.711, however, the received
packets had the PADDING flag set in the RTP header, although the flag was not set when the packets were
transmitted. The PADDING flag indicates the presence of padding octets at the end of the RTP payload, which
cannot be true for G.711. Why this occurred is unknown, but it does not really matter because there is no point in
using the G.711 codec if bandwidth is scarce.
Configuration
1. Specify the number of RTP connections that can be compressed (cache allocation). In interface configuration
mode, the command is ip rtp compression-connections <number>. The default is 32, and each call requires
two connections. The configurable range is 3 to 256 for PPP and HDLC using IOS v11.3 and later; and 3 to
1000 for PPP and HDLC using IOS v12.0(7)T and later. For Frame Relay the value is fixed at 256.
2. The command to turn on compression is ip rtp header-compression in interface configuration mode. It must
be implemented at both ends of the WAN link. For this experiment, when the command was entered into the
router, ip tcp header-compression was also installed automatically. When either command was removed the
other was automatically removed.
Consult Cisco’s documentation for more specific configurations on other types of WAN links (ie, Frame Relay and
ATM) [2 p.5-14, 5-18, 5-26, 5-33] [3]. Configuration for the X330WAN router is very similar to Cisco and well
documented in the X330WAN User Guides.
The following table contains access list guidelines for Avaya media servers and media gateways. Most
connections take place over the S8xxx server’s enterprise interface, which could be a separate interface or combined
with a control network interface. The enterprise interface is…
- Eth4 on S87xx Multi-Connect.
- Typically eth0 on S87xx IP-Connect, but could also be configured as eth4 in some cases.
- Typically eth0 on S8500, but could also vary by configuration.
set port host ? disables channeling/trunking; enables portfast on all or given port(s)
clear port host ? opposite of set port host
set spantree portfast <mod/port> ? enables or disables Spanning Tree fast start feature on given port(s)
show spantree [<mod/port>] displays Spanning Tree and portfast info for all ports or given port(s)
set vlan <vlan id> <mod/port> sets the native vlan (default vlan) for given port(s)
set port auxiliaryvlan <mod/port> <vid> sets the auxiliary vlan for given port(s)
set port auxiliaryvlan <mod/port> none removes auxiliary vlan from given port(s)
show port auxiliaryvlan ? displays auxiliary vlan information
show interfaces status displays settings and status for all ports
show interfaces [fast|gig <mod/port>] displays port(s) status, statistics, and errors at the interface level
clear counters [fast|gig <mod/port>] clears show interfaces counters
show controllers ethernet-controller ? displays port(s) statistics and errors at the controller level
clear controllers ethernet-controllers ? clears show controllers counters
show vlan displays vlan configuration information
set port auto-negotiation <mod/port> ? enables or disables speed/duplex negotiation for given port(s)
set port speed <mod/port> ? sets the speed for given port(s)
set port duplex <mod/port> ? sets the duplex for given port(s)
show port status ? displays settings and status for all ports or given port(s)
show port counters ? displays high level TX and RX statistics for all ports or given port(s)
show ethernet counters ? displays detailed statistics and errors for all ports or given port(s)
clear port counters ? clears statistics and error counters on all ports or given port(s)
set port fast-start <mod/port> ? enables or disables Spanning Tree fast start feature on given port(s)
show port [<mod/port>] displays Spanning Tree and fast start info for all ports or given port(s)
set port vlan <mod/port> <vid> sets the port vlan (default vlan) for given port(s)
set port trunking-format <mod/port> ? sets trunking mode for given port(s)
set port vlan-binding-method <md/pt> ? sets the vlan binding method for given port(s)
show port [<mod/port>] displays trunking and vlan-binding info for all ports or given port(s)
show vlan ? displays vlan configuration information
set port spantree ? <mod/port> enables or disables Spanning Tree on given port (no fast start on
P330)
show spantree [<mod/port>] displays Spanning Tree information for all ports or given port
set port vlan <vid> <mod/port> sets the port vlan (default vlan) for given port(s)
IPSI commands These commands are executed from the IPSI [IPADMIN] prompt.
set port negotiation 1 enable|disable enables or disables IPSI control port (port 1) speed/duplex negotiation
set port speed 1 100MB|10MB sets control port speed
set port duplex 1 full|half sets control port duplex
show port 1 displays control port status and configuration
show control stats displays control port statistics and errors
This rudimentary network configuration is used as a reference point. The objective is to assure high quality
of service to VoIP applications across the congested WAN link.
C-LAN
MedPro
192.168.1.0/24
192.168.2.0/24
From a theoretical standpoint, using separate queues is ideal. When considering the three detriments to IP
telephony – delay, jitter, and loss – audio is more sensitive to delay and jitter, whereas signaling is more sensitive to
loss. This is not to say that audio is not sensitive to loss or that signaling is not sensitive to delay and jitter, but there
are fine tuning points that apply to queuing to optimize it for audio or signaling. From a practical standpoint, in
terms of user experience, these fine points may matter in some cases and not in others.
If the amount of signaling is negligible compared to audio, and if the size of the WAN link is such that
serialization delay is not a factor (typically 768k or greater), then it is feasible to put audio and signaling in the same
priority queue, as long as the queue is large enough to sustain both. In this case the larger signaling packets do not
disrupt audio flow, because the serialization delay is low and because there are so few signaling packets relative to
audio packets. Also, packet loss should not be an issue because the queue should be large enough to sustain both
audio and signaling.
Suppose, however, that the ratio of signaling to audio is much greater – perhaps nearly 1:1. This would be
possible in a remote office where all the signaling goes to a main office but most of the audio is local. Suppose also
that the WAN link is relatively small (typically less than 768k) and serialization delay is a factor. In this case a large
signaling packet entering the priority queue could delay audio packets, and even induce packet loss if the WAN link,
and thus the priority queue, are small enough. It would be advisable in this case to use separate queues, optimized
for the different characteristics of audio and signaling. As stated previously, the X330WAN router and G350/G250
integrated routers are optimized to use separate queues for audio (DSCP 46) and signaling (DSCP 34 or 41).
The preceding paragraphs are generalizations, and are not meant to imply a firm set of rules. Queuing is
very complex, and implementations vary among manufacturers. The explanations given here are intended to give
the reader a starting point. Testing with live traffic and real equipment, coupled with some trial and error, will
ultimately dictate the optimum configuration for a given production environment.
Example 3
Suppose that C-LANs 192.168.1.10 and .11 cannot mark their traffic (pre-Communication Manager system). This
set of configurations is applied only to the left router.
access-list 101 permit ip host 192.168.1.10 192.168.2.0 0.0.0.255
access-list 101 permit ip host 192.168.1.11 192.168.2.0 0.0.0.255
Access list 101 permits any IP traffic from the two C-LANs to the 192.168.2.0/24 network.
There is an implicit deny any at the end of this access list.
class-map match-any untaggedVoIP create a class map called untaggedVoIP
match access-group 101 packets matching access list 101 are in the class untaggedVoIP
Example 4
This is the same as example 2, but with more restrictions on the traffic. In this example DSCP 46 is used throughout
to simplify the access list. A somewhat matching set of configurations is applied to both routers.
access-list 101 permit ip 192.168.1.0 0.0.0.255 192.168.2.0 0.0.0.255 dscp 46 (left router)
access-list 101 permit ip 192.168.2.0 0.0.0.255 192.168.1.0 0.0.0.255 dscp 46 (right router)
Access list 101 permits any IP traffic that is marked with DSCP 46 between the two VoIP subnets.
There is an implicit deny any at the end of this access list.
class-map match-any VoIP create a class map called VoIP
match access-group 101 only packets matching access list 101 are in the class VoIP; this is
more restrictive than matching any packet with DSCP 46
Q1: How does the IP trunk bypass (aka TDM fallback) feature work, and how should the parameters be set on
the system-parameters ip-options form? How do these settings affect the IP trunk bypass feature?
The system-parameters ip-options form is used to define the thresholds that trigger a fallback to a TDM
trunk, thus bypassing the IP trunk. For this feature to work, the ‘Bypass if IP Threshold Exceeded’ parameter must
be set to ‘y’ in the signaling-group form for an IP trunk, and the correct route pattern must be administered. Simply
stated, a near-end MedPro/MR320 monitors network performance by pinging the far-end C-LAN to measure
network response against the configured thresholds.
One thing to note about the IP trunk bypass feature is that it is not fully supported on the S8300/media-
gateway platform. The VoIP module in the G700 does not behave exactly like the MedPro/MR320 board, and it
cannot perform the ping functions that a MedPro or MR320 performs. The issues with an S8300/media-gateway are
discussed throughout this appendix.
When a high threshold is reached the signaling group goes into bypass state, and a fallback TDM trunk is
utilized. When the corresponding low threshold is re-established the signaling group comes back into service, and
the IP trunk is utilized. Because networks and user preferences vary, there is no single set of optimal thresholds.
This is a feature that must be tested and fine-tuned with each implementation. The parameters are as follows.
- Roundtrip Propagation Delay (ms) High: 400-500ms is a good starting point for this threshold. Many users
begin to notice performance degradation at around 200-250ms one-way delay.
- Roundtrip Propagation Delay (ms) Low: 200-300ms is a good starting point for this threshold. 100-150ms or
less one-way delay typically results in very acceptable audio quality.
- Packet Loss (%) High: 7-10% is a good starting point for this threshold. Avaya Labs testing has shown that
audio quality is acceptable even with 5% packet loss.
- Packet Loss (%) Low: 0-3% is a good starting point for this threshold.
- Ping Test Interval (sec): This is the frequency at which pings are sent out. The lower the interval the better for
measuring network performance. In loads prior to Avaya Communication Manager 2.1 the low limit is 10sec,
which is sufficient for detecting a network outage but not for measuring network performance. As of
Communication Manager 2.1 and MedPro firmware v70 the minimum ping test interval is 1sec, which is
granular enough to gauge network performance. 1-2 sec is a good starting point for this parameter.
- Number of Pings per Measurement Interval: This is the number of pings sent out before delay and loss are
calculated. 10 should be used here for a minimum ping test interval of 10sec, which results in calculations
every 100sec to detect a network outage. As of Communication Manager 2.1 and MedPro firmware v70, 20 to
30 pings at 1-second intervals results in calculations every 20 to 30 seconds, which provides the granularity
required to gauge network performance.
Because pings are used to determine network performance, the IP network should ideally give the pings
(ICMP Echoes and Echo Replies) between MedPros/MR320s and C-LANs the same priority as audio traffic. To
facilitate this it is important to know that the call server can select any MedPro/MR320 in the near-end system’s
network region to originate the pings. Depending on the network, it may be feasible to activate this feature without
deploying any network policies for the pings, especially if the primary concern is to compensate for network outages
and not necessarily for poor performance.
Q2: Besides the IP trunk bypass feature, what other mechanisms are in place to detect an outage or severe
congestion in the IP network, and how long does it take to detect it?
See section 3.5, heading “trunk-group and signaling-group” for details on the LRQ feature that applies to
individual calls placed over an IP trunk. For the IP trunk as a whole the best method is the IP trunk bypass feature.
In addition there is also a Maintenance Function. This function assesses the IP trunk every 15 minutes in a G3r or
Linux platform, and every hour in a G3i platform. Without going into detail, the Maintenance Function determines
whether the signaling group is in service or out of service. It can detect a network outage, but it does not assess
network performance.
The scenario for severe congestion is different. In the case of severe congestion the S8700 detects the
congestion and puts the signaling group in bypass state, the same as with a network outage. It then sends a message
to the S8300 indicating this condition. (This message is also sent in the network outage case, but it doesn’t reach the
far end because of the outage.) The status signaling-group command at the S8700 shows the signaling group in
bypass state. The same command at the S8300 shows the signaling group in far-end bypass state. In this condition
both sides use the fallback TDM trunk until the S8700 puts the signaling group back into service.
Q3: As a follow-up to the previous question, what are the effects of the two sides not detecting the outage at
exactly the same time?
Both sides accept incoming calls on TDM trunks, regardless of the state of IP trunks. So if side A detects
an IP network outage and calls side B via the TDM trunk instead of the IP trunk, side B accepts the call. Side B
continues to attempt using the IP trunk until it detects the outage, at which time it utilizes the TDM trunk for its
outbound calls.
In the case of severe congestion, side A detects the congestion first, goes to bypass state, and starts using
the TDM trunk. This causes side B to go to far-end bypass state and also use the TDM trunk. Eventually side B
detects the congestion and goes to bypass state as well (unless the system is an S8300/media-gateway).
Q4: When the IP network recovers after an outage or severe congestion, do both sides discover this at the same
time and start sending calls over the IP trunk at the same time? If not, what are the effects?
No, as with detecting the failure, detecting the recovery is also independent. But this is usually not a
problem because both sides accept incoming calls on an IP trunk in bypass state. So if side A detects the IP network
recovery first and calls side B while B is still in bypass state, side B accepts the call. However, the same is not true
if side B is in out of service state.
Q5: If the C-LAN or S8300 on one end of the IP trunk fails, does the IP trunk cover to a different C-LAN or
S8300?
No, the IP trunk has fixed termination points. If one of the points fails the IP trunk goes out of service
almost immediately at the local system where the failure occurred. This is especially true for an S8300 because it is
the call server and not just a call signaling board like the C-LAN. At the remote system (the other end of the IP
trunk) the IP trunk eventually goes out of service as follows. The IP trunk bypass feature puts the signaling group in
bypass state (unless the system is an S8300/media-gateway). The Maintenance Function, either at the normal
interval or triggered by a call attempt, puts the signaling group out of service. Depending on which of these occurs
first the signaling group may go into bypass and then out of service, or out of service directly. A way to compensate
Q6: What about a MedPro/MR320 or VoIP module failure at either end of the IP trunk?
The IP trunk is not tied to any given MedPro/MR320 or VoIP module. As long as there is at least one
MedPro/MR320 or VoIP module at each end with available DSP resources, the IP trunk is unaffected by
MedPro/MR320 or VoIP module failures. If all usable Medpros or VoIP modules fail, the IP trunk’s trunk group
members go out of service, but the signaling group stays in service and can be used to send messages between the
two systems. This essentially results in a bypass condition where the TDM trunk is utilized.
When configured properly the stations and media gateways have a list of alternate gatekeepers. They
discover if a C-LAN they are registered with has gone down, and re-home to a different C-LAN. If the C-LAN
failure occurs during an active call, the H.323 and H.248 link bounce recovery features preserve active calls on
stations and media gateways, respectively.
Q8: How is call processing affected in general by a MedPro/MR320 or VoIP module outage?
The call server knows when a MedPro/MR320 or VoIP module has gone out of service and stops directing
calls to that device. As long as there are sufficient MedPros/MR320s or VoIP modules to compensate for the
outage, there is no adverse effect. If there is an outage during an active call, and that call is going through the
affected MedPro/MR320 or VoIP module, that call loses audio. Avaya is studying the concept of redirecting an
active call to a different MedPro/MR320 or VoIP module in this type of failure.
If the IP trunk outage is the result of a C-LAN/S8300 failure, direct IP-IP calls remain up until one of the IP
phone goes on hook. If the IP trunk outage is the result of a MedPro/MR320/VoIP failure, existing calls are affected
as previously described. If the IP trunk outage is the result of the IP network going down, the audio is lost on active
calls, and new calls are routed over the fallback TDM trunk if one is administered.
[2] Cisco Systems, Inc., “Cisco IP Telephony QoS Design Guide,” www.cisco.com, Customer Order Number:
DOC-7811549=, Copyright 2001.
[3] Cisco Systems, Inc., “Configuring Compressed Real-Time Protocol,” www.cisco.com, July 2002.
[4] Cisco Systems, Inc., “Troubleshooting Cisco Catalyst Switches to Network Interface Card (NIC) Compatibility
Issues,” www.cisco.com, July 2002.
[5] Cisco Systems, Inc., “Understanding Compression (Including cRTP) and Quality of Service,” www.cisco.com,
July 2002.
[6] IEEE, Inc., “802.1Q: IEEE Standard for Local and Metropolitan Area Networks: Virtual Bridged Local Area
Networks,” www.iee.org, December 8, 1998.
[7] IETF, “RFC 2508: Compressing IP/UDP/RTP Headers for Low-Speed Serial Links,” www.ietf.org,
February 1999.