SAN Assignment Final Report
SAN Assignment Final Report
SAN Assignment Final Report
BELAGAVI
An ASSIGNMENT REPORT ON
By
PRATHUSHA K A 4AL18CS061
Submitted to
Mrs. Shruthi Shetty J
Assistant Professor
CERTIFICATE
This is to certify that, assignment work for the subject “STORAGE AREA
session have been incorporated in the report & scored ___________ Marks out
ASSIGNMENT 1
Module 1: Welcome!
A 5G world has virtualization at its core. As the number of out connected devices mushrooms
from the hundreds of millions to the tens of billions, data centers are relying more and more on
virtualized infrastructure to handle the tsunami of data that we're producing and consuming. And
it's not just data centers: the fact that 100% of the Fortune 100 companies use virtualization (and
VMware virtualization technology at that), tells its own story.
In the Software-Defined Data Center (SDDC), compute, networking, and storage infrastructure
is virtualized so that resources can be pooled and used more efficiently, less expensively, and
faster. Real strides have been made in server (compute) virtualization and are increasingly being
seen with storage virtualization.
The efficiencies gained from them, however, have been limited to a certain extent by legacy, (i.e,
traditional, non-virtual) network infrastructure that's still reliant on physical hardware and mainly
manual processes. While an organization's virtualized compute and storage may be dynamic,
agile, and flexible, its legacy networking just can't keep up. And an infrastructure or organization
that can't keep up often gets left behind.
Network virtualization enables the speed, mobility, and security, needed in a 5G world.
Infrastructure can be made ready for new applications or be changed in minutes, rather than days
or weeks. Apps and workloads are no longer restricted to individual physical subnets, neither are
switches, routers, firewalls, etc. The security focus moves from simply protecting the perimeter
(the outside surface) of a data center's infrastructure to providing the ability to give each virtual
machine and virtual network its own firewall, shifting the focus to the inside perimeter of the data
center and reducing the attack surface. In addition, virtual networks are isolated and segmented
from each other and from the underlying physical infrastructure so that threats cannot be spread
if they do get in.
Network virtualization extends these features and many others to the cloud as well, a critical
factor to the 81% of enterprises that now use multiple cloud deployment models.
Module 1: Welcome!
In one sense, it’s easy enough to know what software-defined storage (SDS) is because the storage
industry itself provides us with a definition. The Storage Networking Industry Association
(www.snia.org), the global trade association, describes SDS as “Virtualized storage with a service
management interface”. (An “interface” is a boundary through which separate systems interact
with one another.) It is important to note that the terms storage virtualization and software-defined
storage are not the same and that storage virtualization is only a piece of the whole SDS stack.
SNIA goes on to say that SDS “must include:
“Automation – Simplified management that reduces the cost of maintaining the storage
infrastructure.”
“Standard Interfaces – APIs for the management, provisioning, and maintenance of storage
devices and services.” (An API is a set of tools that allows one set of software components to
interact with another – we will be discussing these later in the course.)
“Virtualized Data Path – Block, File and/or Object interfaces that support applications written to
these interfaces.” (A data path is a route taken by blocks, files and objects to and from storage
devices.)
“Scalability – Seamless ability to scale (the storage infrastructure without disruption to the
specified availability or performance.” (To scale is to increase/decrease the size of your storage.)
“Transparency – The ability for storage consumers to monitor and manage their own storage
consumption against available resources and costs.”
Department Of Computer Science and Engineering, AIET, Mijar Page 3
STORAGE AREA NETWORKS (18CS822)
But to more fully understand the purpose of SDS, it’s necessary to understand at least some of
the limitations of the storage technologies that existed before it. Limitations such as the
inflexibility and high cost of proprietary hardware – hardware manufactured by one vendor that
cannot be used with hardware manufactured by another vendor. An organization using such
hardware that needs a new feature may need to have the feature custom-made – at more expense.
And increasing the capacity of proprietary hardware by adding more disks is, of course, costly.
Another limitation is how complicated storage technologies have been. RAID levels, striping,
mirroring. Continuous data protection, asynchronous replication, data deduplication. LUNs,
volumes, datastores. Storage administrators have had to know all this (and much, much more)
while keeping up-to-date with the latest developments in the field.
Against this backdrop, it’s easy to see the advantages of software-defined storage. SDS separates
storage software from the hardware it’s managing. Storage is no longer locked into proprietary
hardware because SDS works with “commodity” hardware – that’s any industry-standard or X86
system. Organizations can, therefore, buy lower-cost hardware that genuinely works for them. If
they want to upgrade the software independently of the hardware, they can. SDS helps manage
all the different types of storage device, by adding a software layer that controls all of input and
output operations from the different storage devices.
Module 1: Welcome!
Before beginning to understand how virtualization works and its benefits, you first have to
understand the concept of efficiency. To be efficient means to get the optimum (maximum)
benefit from the resources being consumed. For example, you may be familiar with the term "fuel
efficient" being used to describe cars. If you are selecting a vehicle to use for your daily commute
from the choice of a large sport utility vehicle that gets 10 miles per gallon or a compact car that
gets 45 miles per gallon, which would be the more efficient choice? The compact car would be
the more efficient choice since it provides more miles of use per gallon of fuel (the resource being
consumed).
Although technology has brought us very far, it is not always very efficient. If a large, expensive
computer system is using only a fraction of its capabilities to complete small tasks, one at a time
or sometimes sitting idle, it is not being used efficiently. Just as if you had a load of laundry to
wash, it certainly would not be efficient to wash one piece of clothing at a time. To get the laundry
done in a more efficient way, the solution would not be to purchase more washing machines (too
costly and impractical) but to make more efficient use of the washing machine by washing
multiple items at once. The same applies to computing, virtualization can be the answer to
improving efficiency by enabling a computer to accomplish more than one task at a time.
A key benefit of virtualization is the ability to create multiple copies from an original device. If
humans had the capability to be virtualized it would mean that you could create copies of yourself
in a virtual environment and those copies would be able to function very much like yourself. The
only difference is that they would only exist in the environment they were created in and not in
the "real" tangible world. Obviously, this technology is not available for us humans right now
but, due to virtualization technology, it is a reality for computers. We will explore how this is
possible later in this course but for now, we will introduce you to the two crucial parts of a
machine that are necessary to use virtualization: hardware and software.
ASSIGNMENT 2
2.1 PROBLEM STATEMENT
Seminar on iSCSI
1 iSCSI Protocol Stack
2 iSCSI PDU
3 iSCSI Discovery
iSCSI protocol layers and depicts the encapsulation order of the SCSI commands for their
delivery through a physical carrier.
SCSI is the command protocol that works at the application layer of the Open System
Interconnection (OSI) model. The initiators and targets use SCSI commands and responses to talk
to each other. The SCSI command descriptor blocks, data, and status messages are encapsulated
into TCP/IP and transmitted across the network between the initiators and targets.
iSCSI is the session-layer protocol that initiates a reliable session between devices that recognize
SCSI commands and TCP/IP. The iSCSI session-layer interface is responsible for handling login,
authentication, target discovery, and session management. TCP is used with iSCSI at the transport
layer to provide reliable transmission.
TCP controls message fl ow, windowing, error recovery, and retransmission. It relies upon the
network layer of the OSI model to provide global addressing and connectivity. The Layer 2
protocols at the data link layer of this model enable node-to-node communication through a
physical network.
works on top of the Transport Control Protocol (TCP) and allows the SCSI command to be sent
end-to-end over local-area networks (LANs), wide-area networks (WANs) or the internet.
A protocol data unit (PDU) is the basic “information unit” in the iSCSI environment. The iSCSI
initiators and targets communicate with each other using iSCSI PDUs. This communication
includes establishing iSCSI connections and iSCSI sessions, performing iSCSI discovery,
sending SCSI commands and data, and receiving SCSI status. All iSCSI PDUs contain one or
more header segments followed by zero or more data segments.
The PDU is then encapsulated into an IP packet to facilitate the transport. A PDU includes the
components The IP header provides packet-routing information to move the packet across a
network. The TCP header contains the information required to guarantee the packet delivery to
the target. The iSCSI header (basic header segment) describes how to extract SCSI commands
and data for the target. iSCSI adds an optional CRC, known as the digest, to ensure datagram
integrity. This is in addition to TCP checksum and Ethernet CRC. The header and the data digests
are optionally used in the PDU to validate integrity and data placement.
iSCSI PDU does not correspond in a 1:1 relationship with an IP packet. Depending on its size, an
iSCSI PDU can span an IP packet or even coexist with another PDU in the same packet. To
achieve the 1:1 relationship between the IP packet and the iSCSI PDU, the maximum
transmission unit (MTU) size of the IP packet is modifi ed. This eliminates fragmentation of the
IP packet, which improves the transmission efficiency
An initiator must discover the location of its targets on the network and the names of the targets
available to it before it can establish a session.
This discovery can take place in two ways: SendTargets discovery or internet Storage Name
Service (iSNS). In SendTargets discovery, the initiator is manually confi gured with the target’s
network portal to establish a discovery session. The initiator issues the SendTargets command,
and the target network portal responds with the names and addresses of the targets available to the
host.
iSNS (see Figure 6-6) enables automatic discovery of iSCSI devices on an IP network. The
initiators and targets can be confi gured to automatically register themselves with the iSNS server.
Whenever an initiator wants to know the targets that it can access, it can query the iSNS server for
a list of available targets. The discovery can also take place by using service locationprotocol
(SLP).it is commonly used.