When it comes to storage technologies and techniques that get a lot of press and industry coverage, there are always some that will be overrated or hyped. Not to long ago, I would have put information life cycle management (ILM) and continuous data protection (CDP) on the list of overhyped, under-delivered technologies and techniques. However, especially in the case of CDP, the market and industry are becoming aware that CDP is really a functionality that needs to exist within other data protection technologies, as part of solutions, as opposed to a standalone point product. In the case of ILM, it has been overhyped by to many vendors big and small; it has lost its context and meaning, kind of like supplier relationship management (SRM), both of which are considered buzzwords persona non grata these days by many, except those who like to use acronyms and buzzwords liberally.
There is grid which can be overhyped from the standpoint of trying to get a consistent, non vendor, or technology marketing alliance viewpoint. Likewise, in the past I would have put virtualization on the list of overhyped technologies, however there are to many things happening with virtualization in general, as the technologies continue to mature and evolve.
For example, on the storage virtualization front, there is still a lot of hype around the theme of heterogeneous shared everything and pool everything storage virtualization approaches -- which are really storage systems in disguise. The real market adoption -- by that I mean actual paying customers -- is centered around storage virtualization in the form of virtual tape, disk-based libraries, storage controller based aggregation and heterogeneous data movement solutions. Granted some vendors have had some success with the storage pooling, or with software running on an appliance as a storage system in disguise; perhaps this was implemented under the disguise of avoiding vendor lock-in, as opposed to moving the vendor lock-in to the virtualization appliance.
Don't get me wrong, and I will try not to contradict myself here by spending time talking about "green storage" without adding to the flood of "Green Spam," I am a fan of techniques and solutions that can improve your IT data infrastructure effectiveness. Techniques and associated technologies that can be used to improve energy efficiency, as well as effectiveness of IT service delivery, on a practical basis while helping out the environment, are all good things. In fact, if you want to learn more about power, cooling, environmental and associated IT data infrastructure topics and trends, check out www.greendatastorage.com, hopefully without contributing to the "Greenwash" of IT approaches and hype.
However for now, I will go with the most over-hyped buzzwords, techniques and technologies as those being tied to "Green," as opposed to focusing more on effective power, cooling and data management practices. These include reducing your data footprint as opposed to a focus on CO2 emissions and carbon off-sets as a primary objective, or simply deploying tiered storage without intelligent automated policy-based management or performance effective storage. Let me give you a few common examples of current industry hype around "Green wash" and where the stories are lacking and how they can be enhanced from hype to helpful best practices:
- Shift of focus on data deduplication centered just on backup as opposed to addressing overall data footprint reduction using a combination of techniques and technologies for on-line primary, secondary as well as near-line and off-line archive data. Reducing your data footprint includes leveraging archiving of databases, email and files as well as compression, compaction and single instance storage (SIS).
- Thin provisioning solutions that result in over booking and service disruptions, as opposed to systems that can learn from past usage and activity patterns to avoid over-booking and denied storage access as a means to simply storage management.
- Larger capacity disk drives that consolidate storage at the expense of performance, increasing the server to storage I/O capacity and performance gap resulting missed service level objectives and a higher cost per IOP per watt of energy used. Instead looking at the cost of IOP (or MB/sec) per watt of energy used to delivery a given level of service for on-line primary storage.
Moving forward, I see some great things: continued enhancements with data compression, compaction as well as single instance storage (a.k.a. data deduplication or commonality factoring) being integrated into storage systems and server operating system software. Look beyond basic "green" stories for techniques and technologies that enable more effective usage of resources, including enhancing performance when consolidating as opposed to simple consolidation approaches to avoid further increasing the server to storage I/O performance and capacity gap.
This was first published in August 2007