Interviews

posted on 21 May 2008 03:26

Caringo recently announced an update to its CAStor product, version 2.2. We were intrigued by a couple of points in the release, relating to faster volume recovery and CAStor suitability for cloud computing, and asked Caringo to clarify them. Derek Gascon, Caringo's VP for markeing responded with the answers below:

B&F: CAStor 2.2 looks like a point release that adds faster volume recovery - what does that mean please? Also what are the other new and/or improved capabilities in CAStor 2.2? 

Derek Gascon: "The improved performance for fast volume recovery enables CAStor to recover from a failed node or disk much faster than the previous version. CAStor maintains multiple replicas of each content object (file) on different nodes in the cluster so in the case of a failed node, content is still available and accessible."

"The fast volume recovery means that CAStor will automatically replicate all the content that was on the failed node to other operational nodes in the cluster. CAStor 2.2 is able to recover a volume faster than RAID because the full processing power of all the nodes in the cluster can be utilized for recovery; and this is done while ensuring that application access is not impeded. This is the major feature of note in the 2.2 release other than improvements to administration."

B&F: The release sub-head says 'First CAS technology to enable intelligent cloud storage infrastructure.' What are the features in CAStor 2.2 that enable this? If the features enabling an intelligent cloud storage infrastructure existed in a prior version of CAStor could Caringo explain what they are please? 

Derek Gascon: "The clustered storage features of CAStor have been around since it was first released to the market. However, because of its CAS architecture the product has been perceived of as being only in the archive segment yet it delivers capabilities that extend well beyond that. One of the purposes of this announcement is to overtly position Caringo in the clustered storage segment where we effectively compete against companies such as IBRIX, Isilon and others. Caringo isn't top of mind for customers investigating solutions, but should be a part of their consideration set."

"CAStor delivers the ability to build clusters on low-cost commodity hardware (x86) that can scale capacity and performance seamlessly without impact to data availability or applications. It is a high-performance storage infrastructure that is capable of handling the I/O requirements of small and large files whereas others in the segment are really only best at dealing with large files, e.g. video."

"CAStor is completely self-managing, self-healing and self-balancing, which minimizes administrative effort and cost. Its CAS architecture provides a vast, flat address space that can easily handle billions of objects and virtualizes disks across commodity servers in the cluster to present a single storage pool that can scale from Terabytes to Petabytes. To see how it is being used to support a "cloud" service you can look at http://www.skytunez.com that is serving rich media directly from CAStor. It is not just for archive."

B&F: When Jan van Riel joined Caringo  from EMC (where he was the director of technology for the Centera division) he said that the future of Centera was in doubt. Has any more information come to light about that please?

Derek Gascon: "No, we don't have any further information about the future of EMC's Centera product."

^{[Chris Mellor.]}