IBM has moved rapidly to integrate its Storwize acquisition, announcing the IBM Real-time Compression Appliances for NAS based on the latest IBM multi-core X-servers deployed as a high availability pair. There are two flavors:
- STN6800 supporting up to 8 10GbE connections,
- STN6500 supporting up to 16 1GbE connections.
Wikibon has looked in depth at different data reduction technologies and compared their performance and efficiency. The Real-time Compression technology has one of the highest scores on the Wikibon CORE scale (171), as a result of an impressive time-to-compress and a very low impact on latency. This performance was one of the main reason that Storwize won the Wikibon 2009 CTO award. The interesting thing about the real-time compression technology is that it actually improves performance for storage technology by reducing the amount of data that has to be stored and transmitted.
At the London announcement of the new IBM product, Doug Balog, IBM vice president of storage, said that IBM plans to add real-time data compression technology throughout the IBM storage portfolio eventually. However despite the Storwize name, it won't appear in the V7000 until next year.
Perhaps the most exciting future potential for this technology is that the software reader could be built into servers and terminals and provide a true end-to-end data reduction technology. This would be a secret weapon for IBM as no other technology supplier can claim a data reduction technology that doesn't impact performance. Virtually all vendors offering storage optimization technology for primary storage will admit a meaningful overhead for the reduction process (as measured in elapsed time to reduce data). This includes NetApp Deduplication (formerly ASIS), EMC compression, Ocarina (now Dell) and even Permabit Albireo (which is the fastest deduplication for primary storage on the market).
Action Item: Storwize needed IBM to take its technology (now renamed IBM Real Time Compression) and drive large-scale acceptance. Storage executives should expect that data reduction technologies will become a fundamental technology of all primary storage arrays (block and file) within 18 months.