Arthur Cole spoke with Peter Smails, vice president of marketing, Storwize.
Cole: A lot of people are reluctant to cascade too many data reduction techniques in their storage environments because of the fear of corruption. Storwize has been experimenting with compression and deduplication. What have you found out?
Smails: Our patent-pending real-time compression system is unique in that it is completely transparent to downstream applications, including deduplication. We're not experimenting with compression and deduplication; we have multiple customer deployments where they are deriving significant benefit from leveraging the two technologies. More specifically, the customer reduces their initial data set with real-time compression, which creates an immediate space savings. Subsequently, as they migrate, replicate, and back up, those data sets are obviously then smaller as well. Specifically with dedupe, because there is less data to process, the dedupe process is faster.
Cole: Is there a threshold where too much compression starts to hamper performance of the dedupe system?
Smails: Most importantly, generic compression does not work with dedupe. Only our unique real-time compression is transparent to deduplication. So, with Storwize, there is no threshold. In fact, because we compress in real time, we reduce the amount of data to be deduped, thereby improving the performance and efficiency of the deduplication process.
Cole: Does it make sense, then, to integrate these two technologies on a single platform?
Smails: Deduplication and Storwize compression are certainly complementary technologies. Running them in conjunction, in real time, creates significant customer benefits, including reduced storage requirements throughout the data lifecycle, increased performance and utilization, and associated savings in power, cooling and floor space.