Some things are good ideas from the start, and one of them is the
Storage Area Network (SAN). It has great success potential, especially
potential for delivering on the promise that storage not directly
attached to the server will bring advantages in provisioning and
performance, reliability and availability.
But these very real advantages don’t eliminate the need for IT
maintenance actions to permit the best possible SAN performance and
reliability.
“One of the most significant potential issues, and probably the most
unrecognized, is defragmentation in a SAN storage system. With the
implementation of a SAN in their storage environment, many Windows
Server administrators believe that defragmentation, which they accepted
and dealt with when using DASD storage, has gone away.” This statement,
made by David Chernicoff of Windows IT Pro in his paper, “Maximize
the Performance of Your Windows SAN Infrastructure,” points out a
basic misstep that can be costly to IT administrators.
A SAN is agnostic when it comes to server operating systems. And the SAN
appears to the operating system as locally attached storage. It is
inherent in SAN storage that the operating system isn’t aware of the
type of storage it uses and can’t optimize for a particular storage
model.
This also means that SANs are not built to prevent or correct
fragmentation. When you store information in your SAN, fragmentation at
the file system level creates operating system overhead that interferes
with file retrieval. These events are not caused by the SAN, and it’s
outside the scope of your SAN to do anything about them.
Storage administrators can fail to consider this when they run into
reduced efficiency. When they notice performance problems, many
administrators add additional storage, assuming that a lack of free
space is causing slowdowns. They fail to consider that both performance
and reliability are often affected by the fragmentation
that occurs when Windows Server writes data to any storage ― and that’s
not just DASD or NAS, but also SAN.
This fragmentation often shows up as reduced application performance.
Application response time begins to increase, it takes longer to load
files, and applications loaded from the storage take longer to launch.
Your IT team can receive help desk calls with complaints about a variety
of supposed problems, when the real problem is that fragmentation is
causing data manipulation times to increase to the point where the
delays affect the user.
Since it isn’t a SAN function to manage fragmentation, it’s important to
address the fragmentation at its source. Of course, the ideal solution
to this phenomenon would be to prevent it. That would mean writing to
the SAN using technology that minimizes fragmentation as the data
writes, so your system would be passing fewer, larger disk I/Os to the
SAN. You would thus reduce operating overhead and raise your I/O
efficiency.
Defrag
has been automated for years by Diskeeper Corporation defrag
applications for systems ranging from individual workstations to
enterprise servers. Today Diskeeper® 2010 with IntelliWrite® technology
takes defrag
a step further by preventing fragmentation from happening in the first
place. Its proactive defrag is much easier on the budget than throwing
storage expansion dollars at your SAN performance issues.
