Seriously Storage

IP Storage is a technology that supports block mode data transfer over IP/
Ethernet Networks. IP Storage technology provides for integrating existing
protocols such as SCSI and Fiber Channel, directly with the IP Protocol. The
advantage of this integration is that storage and networking can be merged
seamlessly. It means that the storage will use IP technology to provide for SAN,
storage resource management and provide storage services in the network.

In India, and to a certain extent worldwide, storage has predominantly been
in the form of direct attached storage (DAS). Now, as storage-guzzling operating
systems, applications and solutions get more collaborative, the need for
shareable storage and enhancement of storage as per requirements, is being felt.

The current generation of storage implementations have either gone the NAS
(Network Attached Storage) way or the SAN (Storage Area Network) way depending
on the requirement of the client. In case of NAS devices since it uses file
protocols to service requirements from clients, it is generally termed as
Network filers. Since the NAS uses TCP/ IP, a routable protocol to service
clients, it can address requirements over wide area networks as well as Internet
with proper configuration. SANs have generally been storage on network, enabling
SAN storage to be connected and allocated to several servers, which also means
that the servers can be removed from the shared storage subsystem without
disrupting the operation.

The
advantage of IP-enabling storage is that organizations can now use IP protocol
to manage the storage services within the organization on a separate network
rather than using different protocols for LAN and storage resource.

One has to remember FC technology came in to cater to the storage services to
overcome the then limitation of earlier IP LANS as well as those of SCSI Buses.
In case of Ethernet it was more to do with the transfer rates, rather than the
management and security features. While in case of SCSI it was distance
limitations and addressable devices on SCSI Bus.

With the advent of enhancements in the IP/ Ethernet technologies these
differences are vanishing. IP storage helps and supports building of inexpensive
IP-based storage networks over existing infrastructure, typically over Gigabit
Ethernet networks.

With the proliferation of Gigabit Ethernet Networks in most organizations,
adopting IP storage is a practical choice for LANs, MANs or WANs. This gives the
advantage of remote disk access, synchronous and asynchronous remote mirroring,
Disaster Recovery or even remote backups in most cases.

The current trend in storage solutions market is towards consolidating
(virtualization of) storage along with NAS-SAN convergence. This has come about
with users and organizations demanding cost-effective and simple storage
infrastructure catering to the exponential growth in demand for storage space.

IP storage is one of the key factors which is driving storage technologies
towards this, with cost and technology advantage, the progress towards
unification seems more likely.

Advantages offered by IP as a technology:

• Open standards.
• Available technical man power
• Availability of proven QoS and Security standards and functionalities.
• Cost effective Products and solutions.

From Managing to Management

A heterogeneous mix of storage devices and the use of proprietary
technologies are driving up storage-related expenditure. The rapid growth in the
amount of data used by applications and the constant change driven by the coming
and going of applications and technology have lead us to highly complex storage
environments requiring a great deal of administrator activity and attention.

ISCSI CHAIN: The advantage of SCSI is that several peripherals can be daisy chained to one host adapter, using only one slot in the bus

The result is a storage infrastructure-management challenge that has been
described by some as being so costly that total cost of ownership (TCO) has
replaced initial purchase price as the relevant metric when evaluating storage
solutions.

And all this is resulting in... change.

"Change" requires storage network administrators to create and
maintain spreadsheets with serial numbers, firmware levels and connection
information for Fibre Channel host bus adapters (HBAs), switches and storage
devices. "Change" requires application owners to spend time tracking
their storage utilization, forecasting future data growth and planning for the
necessary application downtime when storage hardware is added or refreshed.
"Change" requires storage administrators to track overall hardware
utilization and plan upgrade activities with an eye toward minimizing
application downtime.

In the storage environment, the costs involved in managing change contribute
directly to an overall TCO that has been described as being so significant, that
the initial purchase price is no longer important.

Answering all of the questions which an administrator faces (see box) with
the help of suitable software, allows us to draw a good picture of the storage
network. The next step is to improve the resolution of this picture:

• How full are the file systems on my hosts and which ones are approaching
  an out-of-space condition?
• Which files, users or groups are leading to out-of-space conditions?
• How full are the databases on my hosts and which ones are more likely to
  have space-allocation problems soon?
• Which user or tablespace is leading to database space-allocation problems?
• How am I using up my allocated file system and database space over time?

In traditional environments, a host system owns storage and
has its own unique volume manager and file system. This model creates a
management stovepipe. Individual system administrators are forced to manage
storage resources on a host-by-host basis. One of the more time-consuming and
frequent changes in the storage environment is the process of growing, shrinking
or moving storage capacity. Companies can evolve to an on-demand operating
environment by consolidating host-based volume management and file system
functions into single points of control in the storage network.

With contributions from editorial advisors -SR
Prasanna Kumar, Senior Product Manager with XSERVE India, and Subram
Natarajan, Senior Solutions Architect, IBM Storage group