At the recent Gartner IOM 2010 conference
in Orlando, Florida, I had the good fortune of listening to a series of interesting topics and meeting some really smart people. As one might have guessed, the bulk of the sessions focused on virtualization and cloud topics. But the one topic that piqued my interested was unrelated to virtualization and cloud - it was deduplication and was hosted by Dave Russell
The intent of the session was to bring forward a some customer examples that were deploying deduplication technologies in their backup and recovery solutions. Most of you that read this blog know that deduplication and data reduction have been a hot topic in the industry. And as you likely know, almost every major vendor out there offers some form of deduplication with its associated benefits.
This session provided us two customers who were willing to talk about their experiences with deduplication and the benefits they've received. One customer is using CommVault and the other is using IBM Tivoli Storage Manager v6 (TSM)
. While both customers showcased the quantified benefits from deduplication, the presentation from the TSM customer went beyond just the benefits of deduplication. The TSM customer revealed their quantified benefits and also identified some of the best practices they developed regarding deduplication.
This particular TSM customer is a large producer of natural gas in the U.S. The customers environment has TSM managing about 1.3 petabytes of data from over 1500+ nodes. Overall, their approach to managing backup storage is do it as efficiently as possible and to reduce the overall amount of backup data under management.
Prior to leveraging TSM deduplication, this customer began with "incremental forever
" and compression. Once TSM v6 was released, they adopted deduplication at the server and client in concert with the other data reduction features provided by TSM.
As they began evaluating their use of deduplication, they had to deal with demands from their internal customers - DBA and Exchange admins like full backups etc. Furthermore, they had to consider how their rate of data change, evaluate retention policies, and ensure that their restore requirements weren't negatively impacted by the use of deduplication.
After significant testing and planning, the customer decided that they would initially deploy deduplication for their Oracle databases and Windows OS and system state backups. The results of using TSM deduplication were impressive ...
Oracle deduplication results - 75% reduction of Oracle backup data after deduplication. This was on 3.8TB of physical space on disk and about 15 TBs of data on tape.
And their results on Windows OS and System State were a whopping 94% ... taking them from 172GB of managed data down to 11.4 GB. In this scenario, the customer leveraged TSM 6.2 client- or source-side deduplication.
Overall, very impressive results. By leveraging the data reduction features within TSM, the customer was able to save by using less tapes library cells, tape drives, and disks.
In the end, the customer stated that TSM data reduction (with deduplication) helped them meet their objectives - efficiently reduce data under management. Furthermore, it allows them to reduce their overall HW costs and meet or improve restore requirements. The last comment the customer made before closing the session was that with all the various TSM data reduction capabilities in production, their job had ultimately gotten simpler now that their environment was running more efficiently ...
This is a fantastic story that I really enjoy sharing. If you are a TSM customer and have benefited from its data reduction technologies, then please give me a shout as I would like to hear your story as well.