GSM Frame Structure-Finalwith Graphs
GSM Frame Structure-Finalwith Graphs
GSM Frame Structure-Finalwith Graphs
1. Introduction:
The MS of a GSM public land mobile network (PLMN) communicates with the
serving & adjacent base stations (BSS) subsystem via the radio interface ,Um, the
Base Trans Receivers Stations (BTS) communicate with the Base Station
Controller (BSC) through the Abis Interface while the BSC communicates with the
Network Switching Sub – System (NSS) through the A interface ( Figure 1
presents the basic architecture of the GSM Network [1] and the interfaces between
the network entities)
PSTN
Um
MS BTS Abis Interface
EIR
F
BTS BSC
A F
MS Um
BTS MSC E G-MSC VLR
C B
D
BTS
Um HLR VLR
BSC G G
MS D
BTS
The GMSC represents the gateway to other networks like public switched telephone network
(PSTN), Integrated services digital network ISDN etc.
The Home Location Register (HLR) stores part of MS’s location information and assists
the mobility management by routing incoming calls to the visitor location register VLR in
charge of the area where the paged MS currently roams. The authentication center AuC
is implemented as a part of HLR and helps in authentication of the MS through its
1
international mobile subscriber identity (IMSI). Stolen, fraudulent or faulty MSs are
identified with the help of equipment identity register (EIR). All radio resources are
controlled by BSC and it is principally in charge of all the processes related to handovers
initiation, frequency hopping, channel allocation, link quality, power budget control,
signaling and broadcast traffic control etc. The MSC’s functions include paging, MS
location updating, handover control etc.[2] The GMSC is often implemented in the same
machines as the MSC. The VLR is always implemented together with a MSC; so the area
under the control of MSC is also the area under control of the VLR.
2. Channel Structure:
A channel corresponds to recurrence of one burst every frame (one burst on a TDMA
frame). It is defined by its frequency and burst position within a TDMA frame. In GSM
there are two types of logical channels:
Ø The traffic Channels are used to transport encoded speech and data information.
Full rate traffic channels TCH/F are defined using a group of 26 TDMA frames
called a 26 frame multi-frame. The 26 frame multi-frame lasts 120ms and the
traffic channels for the downlink and uplink are separated by three bursts. As a
consequence the mobiles will not need to transmit and receive at the same time
which simplifies considerably the electronics of the system ( single synthesizer chip
in MS) and preventing high level transmitted power leakage back to the sensitive
receiver. Half rate traffic (TCH/H) double the capacity of the system are also
grouped in a 26 frame multi-frame. The net bit rate, block length, block recurrence
[3] for full rate and half rate traffic channels are 13Kbps, 260 bits, 20ms and 5.6
Kbps, 112 bits, 20 ms. For full rate speech the block is divided into two classes
according to the importance of the bits (182 bits for class I and 78 bits for class II).
For half rate speech, the block is divided into two classes as 95 bits for Class I and
17 bits for class II. The TCH/F consists of one time slot in each TDMA frame i.e.,
one slot every 4.615ms.
Ø The control Channels are used for network management messages(call set up,
control signaling etc.) and some channel maintenance tasks. These can be
subdivided into BCH ( Broadcast Channel ), CCCH ( Common Control Channel),
SDCCH ( Stand Alone Dedicated Control Channel), ACCH ( Associated Control
Channel)
2
The common control channels help to establish the calls from the MS or the network and
are used to allocate an SDCCH to the MS. The SDCCH allocated is used for signaling
between the MS and the network for location updates, authentication of the MS etc. and
is used for subsequent TCH allocation (for data and speech calls). After Traffic Channel is
allocated to the MS, the SDCCH channels are released. Three types of CCCH can be
defined – The Paging Channel ( PCH -Downlink only) is used to alert the MS of an
incoming call; The Random Access Channel (RACH -Uplink only) is used by the MS to
request access to the network i.e. for allotment of an SDCCH. The Access Grant Channel
(AGCH – downlink only), is used by the base station to inform the MS about which
channel i.e. SDCCH should be used. This channel is the answer of a base station to a
RACH from the MS. The SDCCH can share a physical channel with a BCH or CCCH but
not with a TCH. Thus we see from the above that the only channels that are bi-directional
are the associated control channels (FACCH & SACCH). All other control channels are
either for downlink (BCCH, FCH, SCH, PCH, AGCH) or for uplink (RACH) only. The
control channels FCH and SCH are always sent on Time Slot 0 of the BCCH carrier which
for this reason does not follow frequency hopping. The Control Channel BCCH, RACH,
PCH and AGCH must be assigned to the BCCH carrier only on any even numbered time
slot. The SDCCH can be assigned to any carrier and any time slot. This means that
except for SDCCH, FACCH and SACCH, all other control channels have to be on the
BCCH carrier frequency only. The net bit-rate, block length and block recurrence time of
the control channels is summarized [3] below in Table 1.
3
Note: One 51 frame multi-frame lasts 15/ 26 * 8* 51 i.e. 3060/13 ms
One 26 frame multi-frame lasts 15/26*8*26 i.e. 120ms
A Super frame lasts for 6.12 seconds and contains either 51 numbers of TCH-Multi-
frame or 26 numbers of Control-Multi-frame. Hence the duration of Super-frame is
the same for Traffic Channels and Control Channels. One Hyper-frame contains 2K
super-frame and lasts 3hrs 28mins 53.76 seconds. The frame number FN thus can
have 26*51*2048 values from 0 to 2715647. This FN is transmitted by base station
as a part of Synchronization burst which we discuss later. Figure 2 gives the
schematic arrangement of TDMA frames, multi-frames, super-frames and Hyper-
frames. The 26 multi-frame lasts for 120ms which was chosen as a multiple of 20ms
in order to obtain some synchronization with fixed networks, ISDN, in particular. This
leads to the value of TDMA frame as 120/26 and that of one TS as 120/(26*8)=
15/26ms.
4
1 Hyper Frame=2048 Superframes=2,715,648 TDMA 3 hours 28 min.53.76 sec
0 1 2 3 2046 2047
0 1 2 49 50
0 1 25
0 1 2 7
3.2 Bursts:
The physical content of a Time Slot , TS, is called a burst. There are five types
of bursts each having 15/26ms duration and having 156.25 bits. A schematic
representation of burst in power over time presentation is given in Figure 3.
5
Signal 147 bit = 542.8 micro sec. (TCH)
312.2 micro sec. (RACH)
Level
+4 dB +1 dB
-6dB
-1 dB
-30 dB
10 micro sec.
T/micro sec.
10 micro sec.
156.25 bit = 577micro sec.
6
different training patterns is used associated with the so called BS color codes
which assist in identifying the BS’s. The 26 bit training segment is constructed
by a 16 bit Viterbi channel equalizer training pattern surrounded by five quasi-
periodically repeated bits on both sides .Quality of the received signal RXQUAL
is a key parameter for evaluating network performance. RXQUAL is the Bit
Error rate BER derived from the 26 bit midamble from the TDMA burst.
RXQUAL levels characterize speech quality and dropped calls , where 0
indicates the highest quality and 7 the worst . The stealing flag indicates
whether a 57 bit packet actually contains user data ( set to 0) or FACCH
information (set to 1). The autocorrelation function of the eight training
sequences calculated between the central 16 bits and the whole 26 bit
sequence has a central correlation peak surrounded by 5 zero’s on each
side.[4]
T PAYLOAD F TRAINING SEQUENCE F PAYLOAD T G
57 BITS 26 BITS 57 BITS
Figure 5 : Synchronization Burst (156.25 Bits) T : 3 TAIL BITS ; G: 8.25 GUARD BITS
The training sequence of 64 bits is identical for all BTS. The BS sends
synchronization burst on timeslot 0 of the BCCH carrier. The MS sets up its
time base counters after receiving a synch burst by detecting QN (Quarter Bit
Number = 0- 624) counting the quarter bit intervals in burst, BN (Bit Number= 0-
156), TN (Time Slot Number= 0-7) and FN (TDMA frame number= 0- 2715647).
The value of QN is determined from the 64 bit training sequence, the value of
TN is set to 0. QN increments every 12/13 micro seconds; BN is the integer
part of QN/4; TN increments when QN changes from count 624 to 0; FN
increments whenever TN changes from count 7 to 0. The 78 encrypted bits are
decoded to arrive at the 25-SCH control bits. These 25 control bits contain the
PLMN color code and BS color code (BSIC) and the TDMA frame number. FN
is determined by the relation FN=51((T3-T2) mod(26))+T3+51*26*T1, where
T3=(10*T3’)+1; T1,T2,T3’ being contained in the 25-SCH bits [2]
The synch burst is the first burst that the MS needs to demodulate in the
downlink direction.
3.2.3 Access Burst :
T1 SCH SEQUENCE RACH DATA T2 GUARD BAND 68.25 BITS
41 BITS 36 BITS
Figure 6 : Access Burst (156.25 Bits) T1: 8 TAIL BITS ; T2: 3 TAIL BITS
The access burst is used only for the initial access by the MS to the BTS which
applies in two cases:
7
1. For a connection setup from idle state where a CHAN-REQ message is sent
using access burst.
2. For handover when MS sends HND-ACC message. The access burst has
longer guard period of 68.25 bits to ensure that the access burst fits in the
receiver window of a BTS. We must note that the MS has already
synchronized with the network. The BTS determines the actual propagation
delay when the access burst arrives at the BTS and calculates the distance
of an MS from the BTS and provides the offset time as a 6 bit number
(Timing Advance) to the MS which in turn advances its time base over the
range 0-63 bits to transmit its signal earlier to enable the normal burst to fit
in the receiver window of the BTS. The 36 bit contain among other
parameters the encoded 6 bit BSIC (BS Identifier Code) and contains either
a CHAN- REQ or an HND-ACC message. The access burst always starts
with the bit sequence 00111010 followed by 41 bit synchronization
sequence [1] allows the BTS to recognize the access burst. The access
burst arrives at the base station with a time error of twice the propagation
delay compared to the reception window. The access burst is the first burst
that a base station needs to demodulate in the uplink direction. This allows
a maximum cell distance of 35kms. The exact shift between downlink &
uplink as seen by the MS is 3 Burst Period minus TA. [4]
Figure 7: Frequency Correction Burst (156.25 Bits) T: 3 TAIL BIT;G:8.25 GUARD BITS
AlI 148 bits (142+6) are coded with 0. The output of GMSK modulator is a fixed
frequency signal exactly 67.7 KHz above the BCCH carrier frequency [1]. Thus the
MS on receiving this fixed frequency signal fine tunes to the BCCH frequency and
waits for the synch burst to arrive after one TDMA frame i.e. 4.615ms.
Figure 8: Dummy Burst (156.25 Bits) T: 3 TAIL BITS; G:8.25 GUARD BITS
To enable the BCCH frequency to be transmitted with a constant power level,
dummy bursts are inserted into otherwise empty time slots on the BCCH
frequency. The dummy burst are coded with a predefined pseudo random bit
sequence to prevent accidental confusion with frequency correction bursts. A key
difference between BCH and TCH ARFCN is that a BCH ARFCN has continuous
transmission at a constant power level on all time slots , whereas, a TCH ARFCN
has bursted transmission with power levels that can be different in different time
slots .
8
4. Frequencies Available:
The following frequency bands are specified in GSM [3]:
a. Primary Band: 890- 915 (MHz) mobile transmit, base receive
935 – 960 MHz: base transmit, mobile receive
Allowed Frequencies in one direction = 124 with 200khz spacing
b. Extended GSM 900 Band: 880- 915 MHz mobile transmit, base receive
(including standard GSM 900 band) 925-960 MHz: base transmit, mobile receive
Allowed Frequencies in one direction= 174 with 200khz spacing
c. DCS 1800 Band: 1710-1785 MHz: mobile transmit, base receive
1805-1880 MHz: base transmit, mobile receive
Allowed Frequencies in one direction = 374 with 200khz spacing
d. PCS 1900 Band: 1850-1910 MHz: mobile transmit, base receive
1930- 1990 MHz: base transmit, mobile receive
Allowed Frequencies in one direction = 299 with 200khz spacing
For GSM 900 different categories of mobile there are four power classes with the
maximum power class having 8W peak output power and the minimum having
0.8W peak output power. For DCS 1800 there are three power classes of 4W peak
output power, 1W peak output power, and the minimum having 0.25W peak output
power. For PCS 1900 there are three power classes of 2W, 1W and 0.25W peak
output power. Easy formulas to describe the actual frequency of an ARFCN are
(n=ARFCN):[6]
Ø DCS- 1800: Fuplink (n) = 1710 MHz + 0.2 (n- 511) 512 n 885
Fdownlink (n) = Fuplink (n) + 95 MHz
The radio interface of GSM uses slow frequency hopping. The transmission frequency
remains the same during the transmission of a TDMA burst having eight time slots. In
most cases, the emitting and receiving antennas are not within direct line of sight and the
received signal is a sum of a number of copies of one signal with different phases due to
multipath propagation and reflection. The sum of a lot of phase shifted signals with a
random distribution of phases has an envelope following the Rayleigh distribution. The
fading is frequency dependent. With frequency hopping all the bursts containing the parts
of one code word are transmitted on different frequencies and are hence not damaged in
the same way by Rayleigh fading. When the MS moves at high speed, the difference
between its position during the reception of two successive bursts of the same channel
(i.e. 4.615ms) is sufficient to decorrelate Rayleigh fading. In this case, slow frequency
hopping does not help much . However, when the MS is stationary or moves at slow
speeds, SFH allows the transmission to reach the level of performance of high speeds
(around 6.5 dB gain). The second advantage of frequency hopping is “Interferer Diversity”
where due to different hopping sequences of neighboring interfacing cells using the same
frequencies, the quality improves as the received interfering signal follows a different
9
hopping pattern than that of the cell where the MS is receiving the signal. For a set of n
given frequencies, GSM allows 64*n different hopping sequences to be built.[4] They are
described by two parameters, the MAIO (Mobile Allocation Index Offset) which may take
as many values as the number of frequencies in the set and the HSN (Hopping Sequence
Number) which may take 64 different values. Two channels bearing the same HSN but
different MAIO never use the same frequency on the same burst. On the opposite two
channels using the same frequency list and the same TN, but bearing different HSN,
interface I/n th of the bursts. The sequences are pseudo random except for the special
case of HSN=0,where the frequencies are used one after the other in order. Usually
channels in one cell bear the same HSN and different MAIO’s. In distant cells using the
same frequency set, different HSN should be used to gain from interferer diversity. It is
best to avoid HSN=0 which leads to poor interferer diversity, even with non-identical
frequency sets. The BCCH Carrier frequency (Beacon Frequency) is not hopped i.e. the
channels BCCH, SCH, FCH, RACH, AGCH, PCH must use a fixed frequency to ease
initial synchronization acquisition and reduce system complexity. In most applications, a
cell is equipped with exactly as many TRXs as allocated frequencies. In cells of smaller
capacity the operator may choose to let the channels other than those on the beacon
frequency to hop only on as many frequencies as there are TRXs , or, on as many
frequencies as available . A MS transmits (or receives) on a fixed frequency during one
time slot (+577µm) and then must hop before the time slot on the next TDMA frame after
4.615ms.
5. Cycles
T T T T T T T T T T T T S T T T T T T T T T T T T I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
5.2 TCH/H
A TCH/H in time domain is described as one slot every 16 burst periods( two TDMA
Frames) in average.
t1 t2 t1 t2 t1 t2 t1 t2 t1 t2 t1 t2 s1 t1 t2 t1 t2 t1 t2 t1 t2 t1 t2 t1 t2 s2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Figure 10: TCH/H FRAME t1,t2: two half rate TCHs ; s1,s2: their SACCH/Hs
10
5.3 SDCCH
SDCCH are of two types: SDCCH/8 and SDCCH/4
SDCCH /8 are grouped by 8 along with its associated SACCH/C8 to form the
equivalent of a TCH/F & its SACCH/F. SDCCH/4 are grouped by 4 along with
its associated SACCH/C4 and combined with common channels to form an
equivalent of TCH/F and its SACCH/F. All SDCCH follow a cycle of 102*8(two
51-multi-frame) burst periods i.e. eight groups of four slots separated by 4.615ms (8
bursts periods) every 51-multi-frame for SDCCH/8. These are combined with four
groups of 4 slots for SACCH/8 separated by 4.615ms (8 bursts periods) every
51-multi-frame. Thus during the cycle of SDCCH/8 we need two 51 multiframes.
There are maximum 16 different scheduling for MSs in connection with a SDCCH/8.
Similarly SDCCH/4 follows two 51 multiframes cycle [5]. The SDCCH/ 4 can be
combined with common control channels and sent on TS0. Only one SDCCH/4
combination can be defined for each cell.
D0 D1 D2 D3 D4 D5 D6 D7 S0 S1 S2 S3 I I I
D0 D1 D2 D3 D4 D5 D6 D7 S4 S5 S6 S7 I I I
(Two 51 multi frames shown for complete cycle of SACCH )
Figure 11 (51 MULTI-FRAME=235.38 ms)
D0 – D7 : EIGHT SDCCH/8 CHANNELS EACH OF FOUR SLOTS
S0-S7 : EIGHT SACCH/C8 CHANNELS EACH OF FOUR SLOTS ,ASSOCIATED WITH SDCCH/8
I : IDLE FRAMES
f s bcch ccch f s ccch ccch f s ccch ccch f s ccch ccch f s ccch ccch i
f : FCCH (1 time slot) ; s : SCCH ( 1 time slot) ; bcch ( 4 time slots ) ; ccch (4 time slots ) - PCH + AGCH ;
i : IDLE ( 1time slot )
11
A BCCH together with PCH + AGCH uses 40 slots per 51-multi-frame on the same
TN of the beacon frequency. These 40 slots are built into 10 groups of 4, the four
slots of first group are used by BCCH and the remaining nine by PCH + AGCH
when SDCCH is not combined with CCCH. The other combination is that BCCH
with PCH + AGCH uses 16 slots per 51-multi-frame all on the same TN of the
beacon frequency when SDCCH is combined with CCCH . BCCH then uses the
first block of four slots and PCH +AGCH the remaining three. In both cases the
BCCH information can be sent only once every 51- multi-frame i.e. only once every
235.38ms.
12
5.4.5 CBCH
A Cell Broadcast Channel CBCH follows a cycle of 8*51*8 burst periods i.e.
8 numbers of 51-multi-frame. In each multi-frame the CBCH can be seen as
a part of SDCCH/4. There are two combinations possible:
(a) If the common channel configuration is that of case (b) in para 5.4.4 then
the CBCH can use the same time slot 0 and frequency as the common
channels . It then replaces one of the four SDCCH/4s
(b) The CBCH can use TS0 (but not on the beacon frequency), 1,2, or 3; In
this case the MS in idle mode has to listen regularly to bursts of different
time slot numbers. When CBCH is used, the first block of PCH + AGCH in
the 51- multi-frame cannot be used for paging.[4]
It is allowed to stop the termination of the CBCH incase of congestion and
then these resources can be used by SDCCH during such periods. The
CBCH reduces the number of available SDCCH’s.
Table 2
Channels Unused Slots
TCH/F with SACCH/F 1 out of 26
2 numbers of TCH/F with SACCH /H None
8 numbers of SDCCH/8 3 out of 51
FCCH + SCH+BCCH+PCH+AGCH 1 out of 51
In down Link
RACH/F in uplink None
BCCH +PCH+AGCH 11 out of 51
In downlink
BCCH +PCH+ AGCH+ SDCCH/4 3 out of 51
In downlink
RACH/H+ SDCCH/4 none
In uplink
A TRX may combine eight such groups with restrictions on time slots as
discussed earlier.
13
TS0: Downlink: FCCH+SCH+BCCH+PCH+AGCH+SDCCH/4+SACCH/C4
Uplink : RACH/H+ SDCCH/4
TS1 to 7: Downlink : TCH/F + SACCH/F
Uplink: TCH/F + SACCH/F
f : FCCH (1 time slot) ; s : SCCH ( 1 time slot) ; bcch ( 4 time slots ) ; ccch (4 time slots ) - PCH + AGCH ;
i : IDLE ( 1time slot ) ; D0-D3 : SDCCH/4 ; SA0-SA3 :SACCH/C4 ASSOCIATED WITH SDCCH/4 (requires two 51
multi-frames )
T T T T T T T T T T T T S T T T T T T T T T T T T I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
TS 1 TO TS 7 ( DOWNLINK & UPLINK ) T:TRAFFIC ; S: SACCH;I: IDLE
( one 26 multi-frame shown )
Figure 15 :Channel organization for a small capacity cell
f S bcch ccch f s ccch ccch f s ccch ccch f s ccch ccch f S ccch ccch i
TS 0 DOWNLINK ( one 51 multi-frame shown ) of Beacon Frequency
f : FCCH (1 time slot) ; s : SCCH ( 1 time slot) ; bcch ( 4 time slots ) ; ccch (4 time slots ) - PCH + AGCH ;
i : IDLE ( 1time slot )
RACH /F ON ALL 51 FRAMES RACH (ACCESS BURST)
TS O UPLINK ( one 51 multi-frame shown ) of Beacon frequency
D0 D1 D2 D3 D4 D5 D6 D7 S0 S1 S2 S3 I I I
D0 D1 D2 D3 D4 D5 D6 D7 S4 S5 S6 S7
( Two 51 multiframes of a time slot shown for complete cycle of SACCH )
TWO SUCH GROUPS OF 51 MULTI-FRAME ON ANY TWO TS OTHER THAN TS0 OF Beacon Frequency: DOWNLINK
D0 – D7 : EIGHT SDCCH/8 CHANNELS EACH OF FOUR SLOTS
S0-S7 : EIGHT SACCH/C8 CHANNELS EACH OF FOUR SLOTS ,ASSOCIATED WITH SDCCH/8
I : IDLE FRAMES EACH OF ONE SLOT
S5 S6 S7 I I I D0 D1 D2 D3 D4 D5 D6 D7 S0
S1 S2 S3 I I I D0 D1 D2 D3 D4 D5 D6 D7 S4
( Two 51 Multiframes of a TS shown for complete cycle of SACCH )
TWO SUCH GROUPS OF 51 MULTI-FRAME ON ANY TWO TS OTHER THAN TS0 OF Beacon Frequency :UPINK
D0 – D7 : EIGHT SDCCH/8 CHANNELS EACH OF FOUR SLOTS
S0-S7 : EIGHT SACCH/C8 CHANNELS EACH OF FOUR SLOTS ,ASSOCIATED WITH SDCCH/8
I : IDLE FRAMES
T T T T T T T T T T T T S T T T T T T T T T T T T I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Remaining 29 Time Slots ( DOWNLINK & UPLINK ) T:TRAFFIC ; S: SACCH;I: IDLE
14
( one 26 multi-frame shown for single time slot )
f s bcch Ccch f s ccch ccch f s ccch ccch f s ccch ccch f s ccch ccch i
TS 0 DOWNLINK ( one 51 multi-frame shown ) of Beacon Frequency
f : FCCH (1 time slot) ; s : SCCH ( 1 time slot) ; bcch ( 4 time slots ) ; ccch (4 time slots ) : PCH + AGCH ;
i : IDLE ( 1time slot )
RACH /F ON ALL 51 FRAMES RACH (ACCESS BURST)
TS O UPLINK ( one 51 multi-frame shown ) of Beacon frequency
i i bcch ccch i i ccch ccch i i ccch ccch i i ccch ccch i i ccch ccch i
ONE GROUP EACH ON TS 2,TS4 & TS6 DOWNLINK ( one 51 multi-frame shown ) of Beacon Frequency
f : FCCH (1 time slot) ; s : SCCH ( 1 time slot) ; bcch ( 4 time slots ) ; BCCH ; ccch (4 time slots ) : PCH + AGCH ;
i : IDLE ( 1time slot )
D0 D1 D2 D3 D4 D5 D6 D7 S0 S1 S2 S3 I I I
D0 D1 D2 D3 D4 D5 D6 D7 S4 S5 S6 S7
FIVE SUCH GROUPS OF 51 MULTI-FRAME ON ANY FIVE TS OTHER THAN TS0,2,4,6 OF Beacon Frequency:
DOWNLINK
(TWO 51 multi-frame of a TS shown for complete cycle of SACCH shown )
D0 – D7 : EIGHT SDCCH/8 CHANNELS EACH OF FOUR SLOTS
S0-S7 : EIGHT SACCH/C8 CHANNELS EACH OF FOUR SLOTS ,ASSOCIATED WITH SDCCH/8
I : IDLE FRAMES
S5 S6 S7 I I I D0 D1 D2 D3 D4 D5 D6 D7 S0
S1 S2 S3 I I I D0 D1 D2 D3 D4 D5 D6 D7 S4
FIVE SUCH GROUPS OF 51 MULTI-FRAME ON ANY FIVE TS OTHER THAN TS0,2,4,6 OF Beacon Frequency:
UP LINK
T T T T T T T T T T T T S T T T T T T T T T T T T I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Remaining 87 Time Slots ( DOWNLINK & UPLINK ) T:TRAFFIC ; S: SACCH;I: IDLE
( one 26 multi-frame shown for single time slot )
15
An important factor is the average and maximum loads that are expected for
BTS and how the load is shared between signaling and pay load. For cells
having several carriers and with a large amount of expected traffic on
Common Control Channel eg. Paging, channel requests, channel
assignments, the combination B discussed above is most likely to be used.
The signaling needs for mobiles like those for call setup, location updates
etc. are then taken care by the SDCCH’s. [7]
For cells having one or two carriers the combination A is most likely to be
used with SDCCH’s combined with Common Control Channels on time slot
0. Here the paging capacity of the cell is lower as only three paging blocks
are sent as compared to nine in combination B. we must note the position of
the SDCCH’s in the uplink and downlink direction. If the base station
commands the MS to authenticate itself the response can be sent only 15
frames later (i.e after 15*4.615ms). Thus the command response cycle is
reduced to one multi-frame. If the base station manages a huge amount of
transreceivers it is probable that the number of Common Control Channels
provided by combination B is not enough to handle the work and in such
cases combination C is preferred wherein additional Common Control
Channels are allotted. The CBCH if used is always mapped on to the
second subslot of SDCCH i.e. on TS0 of combination A, & on SDCCH time
slots of combination B & C.
16
network AUTH-REQ is sent and stops when
AUTH-RSP/AUTH-REJ is received
by the network
Identification Timer T3270 in the Maximum value 12 sec, starts when On expiry releases SDCCH
network IDENT-REQ is sent and stops when
IDENT-RSP is received
Timer T3250 in the Maximum value 12 sec, starts when On expiry releases SDCCH
network TMSI-REAL-CMD is sent and stops
when TMSI-REAL-COM is received
Table 4
17
No. of SDCCH type Number of Number Capacity SDCCH in Number of TCH SDCCH/TCH
TRX SDCCH of Erlangs TCH Capacity ratio
Sub SDCCH Erlangs
Channels Sub
without Channel
CBCH with
CBCH
Without With CBCH Without With CBCH
CBCH CBCH
1 SDCCH/4 4 3 0.8694 0.4555 7 2.935 29.62% 15.51%
( Combination A) (0.7012) (0.3490) (2.501) (28.03%) (13.95%)
SDCCH/8
(When paging signaling 8 7 2.730 2.158 6 2.276 119.47% 94.81%
load is higher and (2.4037) (1.8778) (1.909) (125.91%) (98.36%)
SDCCH/8 are
configured on other
than TS0)
2 SDCCH/4 4 3 0.8694 0.4555 15 9.0096 9.64% 5.055%
(On TS0) (0.7012) (0.349) (8.108) (8.64%) (4.31%)
The SDCCH/TCH ratio given in the above table indicates the SDCCH configuration
required for a given Grade of Service of the TCH .
It is important to note that for TCH GOS 1% and 2% the ratio of SDCCH/TCH remains
about the same i.e. the same table applies to different values of GOS for TCH. This
means that the SDCCH configuration does not depend on the GOS of TCH and instead
only on the relationship between TCH-GOS and SDCCH-GOS The same is shown in
Figure 18
18
The above table also tells us that the reduction in SDCCH/TCH ratio due to use of CBCH
is maximum for smaller cells i.e. for smaller cells whenever CBCH is used the SDCCH
resources are more severely constrained .
35.00%
% of SDCCH / TCH Load
30.00%
25.00%
20.00%
15.00%
10.00%
5.00%
0.00%
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
No. of TRX
150.00%
100.00%
Load
50.00%
0.00%
0 1 2 3 4 5
No. of TRX
Without CBCH 1 % GOS With CBCH 1 % GOS
Without CBCH 2 % GOS With CBCH 2 % GOS
Figure 18:- Ratio of SDCCH / TCH Loads for Static allotment of SDCCH
19
Estimated TCH load= 20 mE/subscriber
SDCCH/TCH ratio= 5/20 = 25%
From the above table we can select the configuration which gives SDCCH/TCH ratio of at
least 25%. This leads us to the combination: 2*SDCCH/8. However if the paging load is
less the combination SDCCH/4 + SDCCH/8 can also be used as it is +25%.
The SDCCH/TCH ratio depends on parameter setting , subscriber behavior, size of
location area and service provided in the network. SDCCH traffic in mErlang per
subscriber for each type of procedure ( location update, IMSI attach/detach, Periodic
registration, Call set up etc )can be calculated by multiplying, for each type of procedure,
the number of performances per busy hour by holding time of the channel (in sec) and
dividing the result by 3.6 . A margin for traffic peaks of 15% can be added to the
estimated SDCCH load . Contributions from each procedure added together give the total
SDCCH load per subscriber .
SDCCH configuration when TCH is allotted for signaling when all SDCCHs are occupied
with TCH GOS 2% and 1% can be selected from Table 5 below. It has been assumed
that the limit capacity is reached when 0.5 Erlang of signaling traffic is served by the TCH.
Table 5
No. of SDCCH type Number of Number Capacity SDCCH Number of TCH SDCCH/TCH
TRX SDCCH Sub of TCH Capacity ratio
Channels SDCCH E
without Sub
CBCH Channel
with
CBCH
Without With CBCH Without With CBCH
CBCH (Er) (Er) CBCH
1 SDCCH/4 4 3 2.8 2.0 7 2.93 115.22% 82.3%
Combination A
20
Thus we see that using the Immediate Assignment Command of TCH when all SDCCH’s
are busy leads to higher SDCCH/TCH ratios. The situation is depicted graphically in
Figure 19 below.
120.00%
100.00%
SDCCH / TCH Load (%)
80.00%
60.00%
40.00%
20.00%
0.00%
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
No. of TRX
SDCCH / TCH Ratio (Static Allotment) without CBCH SDCCH / TCH Ratio (Dynamic Allotment) without CBCH
SDCCH / TCH Ratio (Static Allotment) with CBCH "SDCCH / TCH Ratio (Dynamic Allotment) with CBCH"
SDCCH 8 / TCH Ratio using Static & Dynamic Allotment for 2 % GOS
350.00%
SDCCH / TCH Load (%)
300.00%
250.00%
200.00%
150.00%
100.00%
50.00%
0.00%
0 1 2 3 4 5
No. of TRX
Figure 19
21
(e) Whenever location updates are increased, the demand for SDCCH resources
increases. Dimensioning of the location area also depends on the paging load. A
paging message must be sent to all cells belonging to the LA where the MS is
registered. The BTS broadcasts all incoming paging messages. Too large LA may
lead to a paging load in the BTS that is too high resulting in congestion and lost
pages. The upper boundary of a LA is set by the paging load and the lower
boundary by the location updating load. Smaller LAs means larger number of border
cells in the network and hence larger updating load. The LA border cells should not
be in high mobility areas such as highways etc and instead should be in low
subscriber density areas to reduce the load on SDCCH due to location updates and
number of handovers.
(f) Each paging block can fit up to four page requests i.e., either 2 IMSI paging
requests, or, 4TMSI paging requests or 1IMSI+2TMSI paging requests. If the
number of paging groups (to which an MS belongs) is large the paging time
increases as the time before which the right paging block arrives is longer. If the
number of paging groups in a cell is small than call set up time reduces but the MS
power consumption increases at its paging group arrives more frequently. To save
battery a MS does not monitor all the paging channels in a multi-frame , it only
monitors the paging channel belonging to its paging group depending on the
setting of the cell parameter BS_PA_MFRMS which informs the MS after how
many multi-frames ( ranging from 1 to 9) the same paging group is repeated . This
means that a mobile paging block can occur at intervals ranging from 470 ms to
2.1 seconds .
(g) The paging messages are controlled by timer T3113 which starts [3] when the
paging message is sent by the network. On expiry, the network may repeat paging
message and start T3113 gain. The number of attempts is a network dependent
choice. Time T3113 stops when PAG_RSP message is received by the network. If
there are too many paging messages increases the queuing time at the BTS,
something that leads to an increase of the average time for a paging response.
(h) Paging load is also affected by the strategy followed in paging- whether the second
page, after no response to the paging message in the cells where the ms is
registered, is a local page in the same cells or in all the cells under the same MSC
area as the former reduces the paging load but the latter has a better chance of
successful paging. Paging load is also affected by whether TMSI or IMSI is used
for paging. Use of TMSI reduces paging load but at the same time use of IMSI has
a better chance of successful second paging message. If the paging message is
global (when LA is not known in the VLR) its is recommended that IMSI must be
used.
(i) If IMSI/ attach/detach and periodic location update are successfully and regularly
carried out, paging load is reduced as the network more or less knows the location
of the MS. Timer T3212 controls the periodicity of regular location update. A
shorter time period reduces the paging load but increases the location updating
load i.e. load on SDCCH. The value of timer T3212 can vary from 1 deci hour i.e. 6
minutes to 255 deci hour i.e. 25.5 hours.[1] The initial recommended setting can be
for periodic location update every two hours.
22
(j) The MS’s Down Link Signaling Counter (DSC) is initialized [10] to the integer that
is nearest to the value of 90/BS_PA_MFRMS when the mobile camps on to a cell.
This counter decrements by 1 when a mobile is not able to decode a paging
message and increments by 1 when a mobile successfully decodes a message.
Once the DSC reaches a value of 0, a radio link failure is declared and the mobile
does a cell reselection. BS_PA_MFRMS can have value in the range of 1 to 9
multi-frames, so the DSC will range between 90 and 10 . Thus for a
BS_PA_MFRMS=1 it needs 90 bad consecutive messages ( 90 multi-frames) to
declare a radio failure and for BS_PA_MFRMS =9 it needs 10 such messages (90
multi-frames) Thus in either case a radio link failure is declared after 90 multi-
frames (21s).
(k) Paging Capacity of BTS: The paging capacity depends on all the above factors
viz the dimensioning of control channels, size of LA, type of paging request used,
paging strategy, setting of timer T3113, periodicity of periodic location update i.e.
efficiency of the location updates which reduce paging load.
Ø For the non combined case when SDCCH/8 is used on a separate time slot
resulting in increased paging blocks availability.
[(9-(number of paging blocks per multi-frame reserved for AGCH))0.2354] Paging Blocks/ Second
If no blocks are reserved for AGCH the paging capacity becomes for the
combined case as 3/0.2354 paging blocks /second and for the non
combined case as 9/0.2354 paging blocks /second. In this case, the Access
Grant will work in stealing mode which means that paging blocks are
replaced with Access Grant blocks if required .To calculate the paging
capacity of a BTS, three cases are considered :
Case I - it is assumed that all second pages use IMSI to identify the MS,
and , that typically 25% of the pages of an MS result in a second page.
Case II - it is assumed that all second pages use TMSI to identify the MS,
and , that typically 25% of the pages of an MS result in a second page
Case III - It is assumed that no second pages are sent .
It is assumed that there are no global pages i.e. the VLR is properly
dimensional .
A. Case I : Thus for each mobile terminated call 1.25 paging commands
are issued which contain 1 TMSI and 1/4th IMSI.
The number of paging attempt per paging block is: 4/ (1+2*25%)= 2.66 ,
Paging Attempt/ Paging Block (Paging attempt = 1 TMSI + ¼ IMSI)
(since one fourth IMSI equals one half TMSI)
Thus the maximum paging capacity in the BTS for case (i) above is
SDCCH Combined Case : 2.66*3/(0.2354)= 33.89 paging attempts/second
23
The number of paging commands the BTS can handle hence comes out to=
1.25*33.89=42.36 paging commands/ second when no blocks are reserved
for AGCH.
We may assume that the maximum allowed paging load is 50% of the
maximum paging capacity in the BTS to ensure that no pages are lost due
to paging queue in the BTS being full, and that the BTS is able to retransmit
all the paging requests.[5]
This leads to maximum paging attempt/ second capacity in the BTS as
16.94 paging attempts/ sec and the number of paging commands therefore
comes out to 1.25*16.94=21.17 paging commands/second.
B. Case II : Thus for each mobile terminated call 1.25 paging commands
are issued which contain 1 TMSI and 1/2th TMSI.
The number of paging attempt per paging block is: 4/ (1+25%)= 3.2 , Paging
Attempt/ Paging Block (Paging attempt = 1 TMSI + ¼ IMSI)
Thus the maximum paging capacity in the BTS for case (ii) above is
SDCCH Combined Case : 3.2*3/(0.2354)= 40.78 paging attempts/second
The number of paging commands the BTS can handle hence comes out to=
1.25*40.78=50.97 paging commands/ second when no blocks are reserved
for AGCH
This leads to maximum paging attempt/ second capacity in the BTS as
20.39 paging attempts/ sec and the number of paging commands therefore
comes out to 1.25*20.39=25.48 paging commands/second.
SDCCH Non Combined Case : 3.2*9/(0.2354)= 122.34 paging
attempts/second
The number of paging commands the BTS can handle hence comes out to=
1.25*122.34=152.92 paging commands/ second.
This leads to maximum paging attempt/ second capacity in the BTS as
61.17 paging attempts/ sec and the number of paging commands therefore
comes out to 1.25*61.17=76.46 paging commands/second
Case III : Thus for each mobile terminated call one paging commands are
issued which contain 1 TMSI .
The number of paging attempt per paging block is: 4, Paging Attempt/
Paging Block (Paging attempt = 1 TMSI )
Thus the maximum paging capacity in the BTS for case (ii) above is
SDCCH Combined Case 4*3/(0.2354)= 50.97 paging attempts/second
The number of paging commands the BTS can handle hence comes out to=
50.97 paging commands/ second.
24
This leads to maximum paging attempt/ second capacity in the BTS as
50.97 paging attempts/ sec and the number of paging commands therefore
comes out to 50.97 paging commands/second.
SDCCH Non Combined Case 4*9/(0.2354)=152.93paging attempts/second
The number of paging commands the BTS can handle hence comes out to=
152.93 paging commands/ second when no blocks are reserved for AGCH
This leads to maximum paging attempt/ second capacity in the BTS as
152.93 paging attempts/ sec and the number of paging commands therefore
comes out to 152.93 paging commands/second.
The most important rule is that the maximum paging capacity of a BTS
should not be exceeded. Similar calculations have been carried out for the
cases when one block is reserved for AGCH . A summary of results is
shown as under in Table 6.
Table 6 :
Type of SDCCH Number Paging Paging Capacity
used of paging blocks/
blocks Second
reserved
for AGCH
Maximum Theoretical Maximum Paging Capacity
Paging Capacity
Paging Paging Paging Paging
Attempt/ Commands attempts Commands per
Second per second per second
second assuming 50%
max load
SDCCH/4
We can see from the above that the paging capacity for Case III is the
highest i.e. when no second page is sent . After that comes the paging
capacity when TMSI is used for second page i.e. Case II .The third in terms
of capacity is the case when IMSI is used for second page . If strategy of
Case III is adopted then, the risk of unsuccessful paging increases and in
25
the first case the paging capacity reduces although the pages are more
likely to be successful . Hence the recommended strategy is that of Case
II i.e. a second page is sent using the TMSI in areas where paging load is
large . In this case if MS has the wrong TMSI in the VLR, the page may be
unsuccessful . In areas where paging load is smaller then strategy of case I
is suitable . Also the above table tells us that using AGCH in a stealing
mode increases the paging capacity by about 50% for SDCCH/4 and by
12.5% for SDCCH/8.
Table 6:
TRX Type of No of Maximum Maximum Traffic Paging
SDCCH Blocks Paging paging Load Load as per
Used Reserved Commands Load , as per Erlang
for per second mEr Erlang calculations,
AGCH Table mEr
1 SDCCH/4 0 42.36 12.21 2.935 0.023
SDCCH/4 1 28.24 8.14 2.935 0.023
SDCCH/4 2 14.12 4.07 2.935 0.023
SDCCH/8 0 127.11 36.66 2.276 0.018
SDCCH/8 1 112.98 32.59 2.276 0.018
SDCCH/8 2 98.86 28.52 2.276 0.018
The above table tells us that for 2% GOS of TCH channels , 100% second
pages using IMSI for second pages , the Paging capacity is much greater
than the paging load . Even if we configure our systems for retransmission
of all pages then also the paging capacity exceeds the paging load. Also the
above indicates that even if two blocks are reserved for AGCH the paging
capacity is adequate . Hence it is the requirement of SDCCH resources that
decides the location area dimensioning and the limit due to SDCCH/ TCH
ratio requirements will be reached earlier .The factors that may lead to
increased paging load from that calculated above may be : increased global
pages when the VLR does not know the LA of the MS, and hardware
constraints on the BTS –BSC interface i.e. the capacity of paging queue ,
unsuccessful pages due to wrong IMSI/TMSI, lower settings of timer T3113
requiring more frequent paging etc. All of the these are however a part of
proper functioning of the VLR and hardware planning . The paging capacity
per se is adequate even for the lower call holding times taken in the above
calculations .
8. Conclusions :
27
SDCCH resources is the deciding factor as paging capacity exceeds the
required paging load.
Bibliography
[1] ‘GSM Networks, Protocols, Terminology and Implementation’ by Gunnar
Heine, Artech House Mobile Communications Series .
[2] ‘The Pan European Mobile Radio System ‘ by Hanzo L. and Steele,
European Trans. On Telecomm
[3] ‘Global System of Mobile Communication (GSM) Specification Series
05.01-5.10 ’ , European Telecomunications Standardization Institute ,ETSI
Secretariat, Sophia Antipolis Cedex, France
[4] ‘The GSM System for Mobile Communications’ by Mouly and Pautet,
Published by Cell & Sys.ISBN 2-9507190-0-7
[5]’Principles & Applications of GSM’ by Vijay K. Garg and Joseph E.
Wilkes, Prentice Hall Communications Engineering and Emerging
Technology Series, ISBN 0-13-949124-4
[6]’An Introduction to GSM’ by Redl, Weber and Oliphant, Artech House,
ISBN 0-89006-785-6
[7]’Mobile Radio Communications’ by Raymond Steele, Pentech Press
Publishers and IEEE Press ISBN 0-7803-1102-7
[8]’Overview of Global System for Mobile Communications ‘ by John
Scourias , Web Document http:ccnga.uwaterloo.ca-jscouriaGSM index.html
[9] ‘GSM System Engineering’ by Mehrotra A. , Artech House, Boston.
28