Three Resolutions for Storage 2008 (Part 2 of 2)
Why storage pros should focus on resiliency, manageability, and their "junk" drawer
Last week I began this discussion by explaining why the front office/back office feud between business users and storage professionals must stop. This week I offer the final two recommended resolutions storage professionals should adopt by explaining two key focal points.
Resolution 2: We Will Focus on Resiliency and Manageability
Our second resolution is to henceforth make resiliency and manageability key criteria in tech product selection. Companies become so myopically focused on technology acquisition cost and performance that they have ignored solution attributes such as resiliency and manageability. The first question that needs to be asked of any proposed technology solution is "How will I manage it?" The second question is "What do I do when it breaks?"
It is no longer sufficient for companies to settle for platforms that are managed with platform specific management tools. Cross platform management is essential to enable fewer personnel to manage more capacity, bandwidth, and processing power.
Ideally, infrastructure management tools could be organized by business process. All infrastructure associated with a specific process (e.g., a subset of total infrastructure) should be presented as a coherent whole to simplify monitoring, forecasting, and troubleshooting. Absent this type of management tool, you should strive for a management approach that shows all of the infrastructure associated with a specific application. At a minimum, a horizontal view of infrastructure, enabling you to see all storage regardless of its attachment modality or vendor branding, is superior to one-off management tools tied to specific hardware platforms.
The Storage Networking Industry Association's approach to horizontal management (Storage Management Initiative Specification or SMI-S) is, contrary to hype, still very much a work in progress that has not achieved widespread buy-in among storage vendors. Tek-Tools, Symantec, CA, and a few others are closer to realizing true horizontal storage management and are inching toward application-centric management. While none provides the entire soup-to-nuts solution for managing all the processes associated with storage infrastructure, selecting one of these products today, deploying it, and making it a standard to be considered in all future equipment acquisitions can deliver immediate business benefits, including improved visibility into storage resources and improved automation of routine tasks, with the collective effect of driving soft costs (labor) out of storage.
Therefore, select a management toolset and then refuse to buy any technology that cannot be managed with your chosen management solution. Smart folks tell me that doing so could chase as much as 40 percent of the cost of ownership out of storage and even more out of networks and systems, which are much more commoditized at the product level.
Resiliency is a complicated topic that we have addressed many times in this column -- and will likely visit again in the coming year. Too often, technology platforms are deployed without considering resiliency requirements, with the result that businesses are spending far too much to "bolt on" continuity capabilities that could have been "built in" to the resulting infrastructure at far less cost.
A question every storage vendor must now be forced to answer during product selection is how do we recover when your technology breaks? Don't be surprised if the answer involves a convoluted strategy that requires the purchase of multiple copies of the vendor's wares and separate mirroring processes for each box. As an alternative, consider a "resiliency wrapper" such as CA XOsoft, Neverfail, or comparable software that will enable you, minimally, to aggregate multiple recovery and failover processes into one highly manageable scenario-based recovery solution.
As with resource management tools, select your "recovery wrapper" first, then use it as another criteria for vetting your storage hardware selections. If the integral recovery capabilities on the proposed hardware platform cannot be managed in concert with other mirroring or backup capabilities on other platforms using your wrapper, choose a different hardware platform.
This is only a short term solution, however. Eventually, you should devise an infrastructure design that replicates key resources at alternative locations, thereby building in recoverability as a function of good architecture. That will take time, and should be phased in as part of the natural technology refresh process.
Resolution 3: We Will Organize the Junk Drawer
The time has come. Doing anything truly strategic about security, disaster recovery, capacity management, archiving, lifecycle management, or compliance means achieving a more granular understanding of the business value of your data assets: that requires a data management project.
This project proceeds on a business-process-by-business-process basis. It requires the discovery of "data DNA" by analyzing each business process, the application(s) it uses, the data generated and used by those applications, and the specific infrastructure on which applications and data are hosted. The result is a data model for the organization and a scheme for segregating or classifying data based on business value and such operational characteristics as access patterns and change rates (volatility).
Using information derived from the data management project, you can determine what data can be archived, what data requires special services such as encryption, mirroring, or tape backup, what data should be included in legal or regulatory retention and deletion schemes, and even what data should be stored on what media based on value-to-media cost criteria.
This data model provides the springboard for creating business-centric infrastructure, for customizing IT to meet business needs directly. Until we get serious about data management, we will always find ourselves adapting our business needs to this or that vendor's idea of how technology should serve the company, an ultimately pointless and costly exercise.
That's it. These are the three most important resolutions that smart companies will set for themselves in 2008. If we do a good job, chances are that our IT investment will begin once again returning value in the form of measurable competitive advantage to our companies.
I recently read that business pundit Nicholas Carr was promulgating his old saw once again about the irrelevancy of IT. A few years ago, he hit a chord with his book "Does IT Matter?" by consoling frustrated business readers that their bristling over IT investments was entirely justified. He said that the commoditization of technology was eliminating whatever competitive advantage could be derived from IT because all companies were deploying the same technology. Now, he is offering that all IT departments need to disappear, replaced by a few cloud-based service providers that deliver commodity services to all subscribers.
Outsourcing is another trend that often accompanies an economic downturn -- an option rationalized too often by arguments such as Carr's. Outsourcing has its place. Routine processes (payroll check processing, for example) can be outsourced quite effectively, enabling resources to be reallocated to other business critical activities or problems. However, one thing that cannot be outsourced effectively is a problematic or non-routine activity. Knowing the difference is what separates successful outsourcing strategies from less successful ones.
We will deal with storage outsourcing in a future column. For now, the take away from this article is simple: if we can put our three New Year resolutions into practice in 2008 -- to work on healing the front office/back office rift, to set management and resiliency standards and prioritize them as technology acquisition criteria, and to organize our data junk drawer -- we will be addressing the correct question: how do we make IT matter?
Your feedback is important. Please feel free to e-mail me at firstname.lastname@example.org
Jon William Toigo is chairman of The Data Management Institute, the CEO of data management consulting and research firm Toigo Partners International, as well as a contributing editor to Enterprise Systems and its Storage Strategies columnist. Mr. Toigo is the author of 14 books, including Disaster Recovery Planning, 3rd Edition, and The Holy Grail of Network Storage Management, both from Prentice Hall.