Riverbed extends archiving to Azure, OpenStack clouds
Die, tape arrays, die, die
Riverbed is stretching its Whitewater appliances – which automatically back up data to online storage clouds – to cover more clouds, and increasing the number of programs on servers that can feed data to them.
The Whitewater appliances were initially launched to archive data using IBM's Tivoli Backup Recovery Media Services (BRMS) and Symantec's NetBackup and Backup Exec archiving programs, and were able to act as the go-between linking systems to Amazon's S3 storage cloud or AT&T's Synaptic Storage Service – the latter being built using Whitewater appliances on the other end of the wire as well, as it turns out.
The idea behind Whitewater is to use the same archiving software that data centers are familiar with, but to push it out to cloudy storage and do the WAN optimization and deduplication functions that speed up archiving.
The idea is also to get rid of tape drives and tape arrays, and having to worry about archiving tapes in some vault somewhere – which, no doubt, will make many companies nervous.
Moreover, the Whitewater appliances have enough capacity to hold about 30 days' worth of data (depending on the size you buy) so you can retrieve that data locally if you need to do a recovery and if the data is not too old.
If you do need to recover data from the cloud, however, the Whitewater appliance is going to do it a whole lot faster than a normal Internet link would. Whether or not restoring from the cloud is faster than tape depends on a lot of factors, including connection speed, the number of files, and the size of files. The other thing that Riverbed is trying to peddle is the fact that the combination of cloud storage plus one of its fancy-schmancy appliances is cheaper than buying a disk array and backing up to that.
In April, Riverbed added some backup software to the list supported by the Whitewater appliances, putting EMC NetWorker, Quest vRanger, and CA ARCserve Backup on the server side of the appliance, and the cloud storage service from Nirvanix on the back end as an optional bucket in which to store the data.
With the current round of announcements, Riverbed is putting Microsoft's Azure cloud and Rackspace Hosting's Cloud Files on the cloudy back-ends, as well as any storage cloud based on the OpenStack Object Storage, also known as Swift.
On the front-end, the Whitewater appliances running the 1.3 release of the software embedded inside the devices can now accept data pushed out from CommVault's Simpana for Backup and Recovery, HP's Data Protector, and Veeam's Backup & Replication tools. With these additional programs, says Ray Villeneuve, general manager of the Whitewater product line at Riverbed, the company can cover about 85 per cent of the addressable market.
In addition to working with these new backup/archiving programs, the updated Whitewater appliances can also do direct SQL dumps from Microsoft SQL Server out to cloudy storage, and similarly can work with Oracle Recovery Manager to push Oracle 10g and 11g databases out to the cloud.
Riverbed's virtual and physical Whitewater appliances
Riverbed sells a virtual Whitewater appliance that runs on a VMware ESXi hypervisor on an x86 server, and delivers around 250GB per hour of input from the archiving software. The physical appliances, which are designated by the 510, 710, and 2010 models, range in speed from 400GB to 1TB per hour. Prices range from $7,995 for the entry virtual appliance to $120,000 for the top-end 2010 model.
These new appliances and the Whitewater 1.3 software will be available before the end of 2011. ®