If Disks are Round, Brown and Spinning, Why is Storage So Hard?

Years ago when I first got involved in storage, I couldn’t figure out why storage management was so difficult. In fact, I initially had a hard time even keeping myself busy in my job as a storage administrator. While I was on a storage management team that was extremely knowledgeable, all of my co-workers worked on mainframe storage. As a result, they were of a little or no help in helping me prioritize what tasks I needed to accomplish. In fact, about the only thing we shared in common was that we were responsible for managing storage systems that had round, brown and spinning disks in them.

Granted, things have changed significantly over the past few years and there is a lot more information available about distributed storage management now than there was back then. But for those individuals who are just starting out in storage or trying to figure what is so complicated about managing storage systems that contain disks that are round, brown and spinning, there is no one reason why it is hard. Rather there are more like a thousand little reasons that, when added together, make it difficult. Here are just some of those reasons:

  • Multiple types of disk drives. Storage systems come with two types of disk drives. Though technically there are Fibre Channel (FC), Serial Attached SCSI (SAS), SCSI and Serial ATA (SATA) disk drives, the disk drives with FC, SAS and SCSI are the same kind of disk drive underneath the covers with different interfaces for connectivity. These drives are generally considered enterprise caliber drives with higher reliability ratings and better performance. Conversely SATA drives are those typically found in PCs. These are lower cost drives with higher capacities but lower levels of performance and reliability. Why makes this complex? Enterprise companies can save significant amounts by mix-and-matching these different disk drives within their storage systems to match application requirements which makes these systems more complex.
  • Multiple storage systems from multiple storage vendors. Understanding the different kinds of disk drives and their make-up is only the first, small step towards understanding storage. There are dozens of storage systems providers that each offer one or more storage systems. These storage systems are then further segmented by how a company intends to use them. The ways storage vendors typical segment their storage systems according to what purpose they serve; such as highly available, high performance applications; file servers; archiving and backup.
  • Multiple storage protocols. If acronyms like SAN, DAS, NAS, SCSI, iSCSI, FCoE, CIFS, NFS and FC aren’t readily recognizable to you, much less roll off your tongue, then you know are still a novice in the storage space. However engineers in enterprise organizations deal with all of these storage protocols, know how each of these differ from one another and know what the hidden gotchas are of implementing each of them.
  • Storage virtualization makes storage harder, not easier. I’m a big proponent of storage virtualization and will remain one but not because I believe storage virtualization makes storage easier. Rather, it makes storage harder. To the untrained eye, virtualization makes everything looks easier. But to the knowledgeable storage engineer who has to configure and manage the infrastructure, all it does is make the moving parts harder to find when troubles occur. The reason I remain a firm proponent of storage virtualization is it is an enabler for those who understand their storage infrastructure and want to take storage management to a higher level.

Managing distributed storage system has made significant progress in the last 6 or 7 years  – both in terms of becoming easier to manage and having more information available so it can be managed. But users have more choices than ever as to what storage systems they can select, how to configure them and mixing-and-matching applications that run on them. Because of that, optimally managing round, brown, spinning disks is going to remain a complex problem for the foreseeable future.

image_pdfimage_print
Jerome M. Wendt

About Jerome M. Wendt

President & Founder of DCIG, LLC Jerome Wendt is the President and Founder of DCIG, LLC., an independent storage analyst and consulting firm. Mr. Wendt founded the company in November 2007.

One Comment

  • Alani Kuye says:

    Virtualization is a great concept however there are lots of unseen pitfalls. Same with Cloud computing. We have to be careful not to invest our entire infrastructure into “virtual environments”. I have seen some organizations running “virtual archiving” systems, which is a contradiction of the truth. You simply cannot have a “Virtual Archive”. A glitch and the system goes toast. Just like paper will never be eliminated, there are some areas within the technology infrastructure that simply cannot be virtualized or sent to the “cloud.
    Alani Kuye
    Phantom Data Systems Inc.
    https://www.phantomdatasystems.com

Leave a Reply

Bitnami