EMC Education Services
I’m a routing geek. Not a storage, compute, SONET, web design, and mobile phone geek — a routing geek. But even routing geeks need to know something about the stuff that attaches to the network right? In the spirit of learning something new, I recently picked up (and actually read!) a book about storage.
This book starts from the very basic basics, discussing the concepts surrounding storage at a fundamental level. There’s an introductory chapter, after which the authors move into the data center — really a chapter about basic storage concepts focusing on disc drive operation, computing IOPs, and direct attached storage. The really useful bits here are the discussion around storage performance, and its impact on application performance. The next chapter discusses RAID, including a valuable section on the impact of RAID on performance, followed by a chapter on intelligent storage systems and dynamic provisioning.
The second section of the book dives into storage networking, starting with Fiber Channel. While the chapter provides a good overview of the technologies involved, there was one key piece missing: how does Fiber Channel actually work? I was really looking for a TCP operational data flow description of the protocol’s operation, which would have helped me get my head around why FC operates the way it does, but there’s not one here. Zoning, virtualization, and FC fabrics are discussed, but all of these are a bit difficult to follow without a solid play by play of the protocol’s operation.
Chapter 6 covers FCoE, FCIP, and iSCSI, technologies that help to bring the data center onto a single Ethernet fabric (if you’re not too much into data center design, having a single fabric for storage and network would be a huge boon to flexible and scalable designs). Chapter 7 discusses network attached storage, and chapter 8 object based and unified storage. There’s a really good discussion around data deduplication in chapter 8.
Section three works through backup, archival, and replication. The first chapter in this section, chapter 9, discusses business continuity as a general concept. I considered this chapter to be fairly weak, but business continuity is probably more formal and wide ranging on the network side of the data center than it is on the storage side. The terminology pieces in the three chapters in this section were helpful, however.
Section four is an overview of cloud; it adds little new information to the field. Section five covers security and management from a storage perspective. Security domains, the risk triad, and encrypted storage are covered here. It would have been interesting to have a section on the impact of encryption on data deduplication, as well as on the performance of storage systems in IOPs and other measures, but they weren’t included here.
Overall, there are some bits to really like about this book, and there are a number of places where it seems like the authors were just filling in a checklist of what they thought needed to be covered. I would still recommend this book for network engineers, though I might skim through sections four and five, and focus my attention on sections one and two.