The second new feature of the DL that I would like to highlight this week is even more straight forward than the last: Oracle Filters. These Oracle Filters give the DL a nearly 2:1 capacity advantage over competitive systems when backing up Oracle RMAN data streams.
To understand what the Oracle Filter is, we need to step back a second and understand how deduplication works. When a data stream is sent to a deduplication appliance, to achieve optimal deduplication efficiency, that appliance should apply a filter that is specific to the type of backup application and data type that is being backed up.
If I am using NetWorker, my deduplication appliance should know that I am using NetWorker and use a NetWorker filter to deduplicate the data sent to it. If I am using NetBackup, the same logic holds. Essentially these filters identify meta-data that is embedded in the backup stream by the backup application and strip them out so that they don't confuse the data set.
Now this is not to say that if I don't have a filter, deduplication cannot occur. Just that it will not be as efficient as it could be.
And this is the case with Oracle RMAN. It is possible to backup and deduplicate an Oracle RMAN stream to DL and get acceptable deduplication ratios. However, with DL software v1.1 we have added a specific RMAN filter. This has effectively doubled the amount of deduplication that we expect to see from an RMAN backup stream.
And of at least as great a significance: this is a feature that Data Domain does not have. As I alluded to in a post a couple of weeks ago about deduplication efficiency, this can have a very real impact on the capacity required. If a Data Domain system will require 33 TB of capacity to store a set of Oracle RMAN data, the EMC DL may only require 18 TB.
The EMC DL will offer nearly 2:1 capacity savings over a Data Domain system for Oracle RMAN data.
Finally, note that this is not just an abstract concept and the savings are not theoretical. The values quoted here are based on actual tests conducted with a real Oracle RMAN data set. The configuration is valid if Oracle RMAN is backing up directly to the DL1500/3000 (in NAS mode) or if it transmits the backup data via a backup application (in VTL mode) with NetBackup, NetWorker, and CommVault. (Remember to set filesperset to 1 as well to optimize deduplication.)