Source Side Dedupe

Comments on 'Source Side Dedupe'

• 

  The backup software companies are finally following Avamar's lead and going after source-side deduplication. Most were caught off guard by the huge trend to deduplication of backup data.
  
  Avamar is the leader in source-side dedupe but you should also check out some other software vendors including Symantec (with Puredisk), IBM/Tivoli and CA with Arcserve. I don't know off hand if HP offers source-side dedupe with Data Protector.
  
  Also, Falconstor is a backup specialist that you should look at. They are entirely focused on data protection and have partnered with Arkeia to deliver source side solutions. Veeam is worth a look. Commvault also has source-side de-dupe
  
  You may also want to consider other approaches such as incremental forever and CDP, depending on what problem you're trying to solve.
  
  Can you describe in more detail what issue you're trying to address?
  
  

  
  

  Posted By:David Vellante| Thu Dec 09, 2010 09:20

• 

  Yes, many product are victims of partial de-dupe. They do half the job so they can fly the “de-dupe” flag. Source side de-dupe is one of the three data reduction techniques. The trick though is all three need to be employed.
  
  AIMstor for instance does source side de-dupe. True to definition, it doesn’t transfer data that already exists in the target store. So if machine A has already sent data into the store, there is no reason why machine B should transfer the file into the store. The second major weapon to the armory is the ability to do live change byte transfers. That means auto detect what has changed and transfer it. 20K modification to a 2GB PST file means only a 20K data transfer. Lastly of course, is post de-dupe. Although AIMstor has all three, we found the post-dedupe has limited value once the other two strategies are employed.
  
  Adding intelligent Data Classification to the mix offers more efficiency as well.
  
  http://www.cofio.com/AIMstor-Backup/
  http://www.cofio.com/Data-Deduplication/
  

  
  

  Posted By:Tony Cerqueira| Fri Dec 10, 2010 11:23

• 

  Thanks Tony and David for your comments. We are in the process of adding source side dedupe support for Backup&Recovery solution aimed primarily at laptops and desktops. I think the requirements of source side dedupe in this scenario include the ability to be able to handle thousands of simultaneous connections, handle thousands of clients as well as enable seamless end user restore. We are in the process of evaluating our options between build / license / buy ? Do the solutions mentioned by you PureDisk, Avamar, Cofio etc meet these requirements?

  
  

  Posted By:Giridhar| Sat Dec 11, 2010 02:11

• 

  Some things come immediately to mind that would help:
  
  Real-time Changed Byte Transfer, with ability to cache local changes until connection is re-established (particularly for laptops). This helps avoid scans and heavy trawls of thousands of client file systems every time a backup takes place.
  
  As for handling thousands of clients, ability to easily create large or small machine groups, and to attach policies to the groups helps when you want to change a policy easily, or change the flow of a whole group to a new storage device or a new network link.
  
  Additionally, using a Distributed Node approach (with self managing nodes, even laptops) as opposed to Client/Server, so as not to worry about backup servers or slave / media servers going down and interfering with backup execution.
  
  A UI for users to restore a folder, a file, and to use the restore like an archive is also helpful, especially if the user can do instant search for their data on a specific file based on rich metadata - all from just their backups.
  
  You can reach us via website at:
  
  http://www.cofio.com/Feedback-and-Inquiries/
  or email sales@cofio.com
  

  
  

  Posted By:Tony Cerqueira| Sat Dec 11, 2010 03:01

Post A Comment

You must be logged in to post a comment, please Sign in