Storage Area Networks (SANs) are a great boon to enterprises everywhere. Because a SAN removes network traffic from the production system, network traffic is freed up to accommodate day-to-day operations—themselves a heavy load. SANs generally implement multiple physical disk drives in some form of fault tolerant disk striping (RAID), and do provide a great benefit to an enterprise: because stored data does not reside directly on any of a network's servers, server power is utilized for business applications and network capacity is released to the end user.
Connecting a machine to a SAN has always been a bit of a task—it normally has to be performed manually, and with today’s heterogeneous environments there has to be considerable know-how involved in the machine’s interaction with the SAN. It becomes even more complicated, however, with the advent of virtual machines (VMs)—for each VM, a “relationship” must be established with the SAN. Since VMs can now be created and deleted on-the-fly by the users themselves, automated solutions are now appearing that will allow VMs to be automatically connected. Whether this will be a workable solution or not remains to be seen, but obviously something needs to happen to make this operation efficient.
File fragmentation already negatively affects SAN performance, if not fully addressed with an automatic solution. Physical members in a SAN environment are not read or written to directly by an application, but instead are “seen” by an application and even the OS as one single “logical” drive. When an I/O request is processed by the file system, there are a number of attributes that must be checked which cost valuable system time. If an application has to issue multiple "unnecessary" I/O requests, as in the case of fragmentation, not only is the processor kept busier than needed, but once the I/O request has been issued, the RAID hardware and software must process it and determine to which physical member the I/O request must be directed. When files are fragmented into hundreds, thousands or tens of thousands of fragments (not at all uncommon), there are obviously many more extra I/O requests. Performance slows to a crawl.
With all that must be done to keep a SAN up and running and to ensure all machines and applications are connected, IT personnel cannot afford to be chasing down and addressing symptoms of file fragmentation. Especially with the addition of VMs, there is already enough to do. Fragmentation must be constantly addressed so that is simply eliminated—a task that can only be performed with a fully automatic solution. Such a solution works invisibly, in the background, with no negative impact on system processes and—best of all—no required scheduling by IT personnel.
Don’t let fragmentation add to SAN complexity. Make sure your fragmentation solution allows you to address factors that truly need addressing.
No comments:
Post a Comment