Book Review: "Data Center Storage"

I recently read an advance copy of a book that’s due to be published in mid-February. I know and respect the author, and it’s rare that we get entire books on our favorite subject, so I was anxious to read it. It’s called Data Center Storage: Cost-Effective Strategies, Implementation and Management, by Hubbert Smith (see bio at the end of this post).

This book makes the case that IT organizations spend too much on storage, examines why, and provides practical advice on how to correct that situation.

According to Hubbert, the root causes of over-spending include using direct-attached server storage instead of consolidated storage; using old school performance-optimized designs rather than capacity-optimized storage; relying on single-tier storage rather than tiered storage; backing up to tape in situations where snapshots should be used; and not employing tiered storage SLAs. And, of course, over-spending on storage will accelerate as volumes of incoming data ramp up.

After examining the root causes behind over-spending on storage, Data Center Storage then provides solutions to “deliver more while spending less.”

The author sets out to separate the signal from the noise surrounding data center storage through use cases and coverage of key issues in storage performance, capacity, power and cost. The examples are usually supported with simple “financial scenario A vs. financial scenario B” cost analysis, including hardware, maintenance and operational (people) costs. The book goes deep enough into the technologies to offer IT-business-level understanding and recommends specific improvements for IT managers.

Data Center Storage then covers the building blocks of storage, including hard disk drives (HDDs) and solid-state disk (SSD) drives, in business terms such as IOPS/$ and GB/$, leading the reader to an understanding of the tradeoffs between storage capacity, performance and costs. Data center power is covered to an adequate depth to improve IT-business-level understanding, and at the end of every section the book offers recommendations for improvements -- in the context of financial scenarios -- to assist in the justification of spending on improvements.

Storage consolidation and data protection are also covered and, consistent with other sections in the book, improvements are provided with financial comparison templates.

The next sections of the book build on the storage consolidation theme, introducing storage tiering and tiering-specific Service Level Agreements (SLAs) covering performance, capacity, up-time, growth, RPO/RTO and price. All of this is in support of the book’s central theme: delivery of service levels and fiscal responsibility.

Data Center Storage de-mystifies new, and not-so-new, technologies such as virtualization, replication, snapshots, thin provisioning, and unified storage (SAN/NAS). The book provides insight into where these technologies have a payoff and where they have limited payoff.

The final sections of the book build on storage consolidation, tiering, reliability and SLA approaches to guide your organization in taking advantage of managed hosting and cloud services, including project planning and service provider vetting and SLAs, all of which is supported with the “financial scenario A vs. financial scenario B” concept.

Data Center Storage closes with insights into roadmap creation, project management and financial justification drawn from the author’s experience in technology company processes, which are equally applicable to IT.

This book is not for readers looking for a technology-only book -- those that exhaustively describe the inner workings of all the various RAID levels and the intricacies of Fibre Channel zoning and similar technical deep-dive topics.

This book is for IT storage professionals who manage storage systems, especially those lobbying for storage improvements to meet “spend less and deliver more” directives. Data Center Storage improves the possibility of fruitful interactions with the folks that control the budget and spending on storage.

I’d also recommend the book for IT business professionals that are tired of the marketing noise and weary of hearing about all the solutions in search of problems. The book provides a business-level understanding of storage use cases and associated costs, and should lead to more informed spending decisions.

Most storage books cover technology, and not business, issues. In contrast, Data Center Storage offers a unique and useful approach to business-meets-storage-technology. And, as the fundamental components of data center storage (performance, capacity, reliability/up-time, power and price) do not change significantly over time, this book could be an often-referenced source for any IT department -- small, medium or large.

For more info or to place an advance order, visit the Data Center Storage page on Amazon.com.

And here’s a little background on the author:

Hubbert Smith contributed to early versions of several disruptive technologies, including network-attached storage (NAS), server fail-over and storage replication. He also pioneered the creation and development of a new product category -- capacity optimized enterprise disk drives (aka Raid Edition) – which is used in enterprise storage and surveillance systems. Capacity-optimized drives now make up roughly one-third of all enterprise-class disk drives.

Smith contributed to the development of Serial ATA (SATA) industry standards, specifically those that enabled SATA to be used in enterprise storage systems. He is a patent holder and a published author (Serial ATA Architectures and Applications - Intel Press).

He holds a BS degree in Electrical Engineering from Auburn University and is currently Director of Product Management at LSI.