Cloud storage brings cost-effective offsite backup and retrieval capabilities to small and midsize businesses, but using cloud storage as an external disk is not as effective as it seems, particularly as the data set grows. It takes time to send and retrieve files to a cloud storage provider and often means using another set of tools to do so. Riverbed's Whitewater appliance makes the cloud storage appear as a backup target and balances the competing needs of speedy read/writes long-term bulk storage. Using a combination of local file caching and deduplication with cloud storage replication, IT gets the best of both worlds.
Small and midsize organizations frequently face a challenge getting their backup data offsite on a regular basis. The traditional solution has been to periodically send backup tapes to a records warehouse run by Iron Mountain, Recall or the like. That creates a lot of work for backup administrators, as they have to create and manage additional jobs to duplicate backup data to tapes for offsite storage, then box up the tapes and deal with the courier. The result is that most midsize companies send tapes offsite only once a week, leaving their data vulnerable.
Organizations with more than one data center can install deduplicating backup appliances and replicate their backup data from one data center to another while deduplicating over the WAN, assuming, of course, that they're using the same brand appliance in each data center. Organizations with just one data center have to either stick to tape or use an Internet-based backup service.
Online backup services like LiveVault and eVault have been available for 15 years or more, but many organizations have been uncomfortable converting from the backup software they've spent months getting to work properly and handing their backup process to a third party lock, stock and barrel. As more general-purpose cloud storage providers like Amazon S3 and Nirvanix have developed, IT professionals have looked for the option to use their existing backup software to send their backup data offsite to a cloud provider. Add to all this the fact that most cloud storage providers charge significantly less than their online backup brethren to store each gigabyte for a month, and cloud backup starts looking very attractive.
Backup applications have started adding the ability to duplicate backup data to cloud providers, but most don't deduplicate your data before sending it to the cloud. Since you're going to be paying the cloud provider of your choice for each and every gigabyte of data they store for you each month, deduplication would definitely be a good idea.
Riverbed's Whitewater appliances act as gateways between your backup applications and the cloud provider of your choice. They look to your backup application like network-attached storage, accepting data via SMB and/or NFS. Once you send your data to the Whitewater appliance, it compresses, deduplicates and encrypts your data using AES and sends it off to your cloud provider. Unlike Riverbed's network deduplication products, Whitewater does not re-hydrate your data at the cloud storage provider.
The Whitewater product line ranges from a virtual appliance starting at a list price of $7,995 to three physical appliances with 2, 4 and 8 Tbytes of usable disk space, respectively, starting at list price of $23,995. Since the disk in the appliance is just a cache, there's no technical limit on how much data you can store on a Whitewater appliance. Whitewater appliances can store their data on most public cloud providers, including AT&T Synaptic Storage, Amazon S3, Microsoft Windows Azure, Rackspace Cloud files and Nirvanix. You can also use them to front-end a private object storage infrastructure build on EMC's Atmos or Openstack Swift. Howard Marks is founder and chief scientist at Deepstorage LLC, a storage consultancy and independent test lab based in Santa Fe, N.M. and concentrating on storage and data center networking. In more than 25 years of consulting, Marks has designed and implemented storage ... View Full Bio