Interest seems to be picking up in the mid-range unified storage space lately. In the past month or so, we've seen Iomega and NetGear move up market with 12-drive systems with enough intestinal fortitude to host Exchange and VMware via iSCSI. Moving up-market, Xiotech, Compellent and Nimbus have all announced higher-end NAS/block systems, and rumors have been floating around that EMC is about to merge their Clariion and Celerra lines.
Enterprise players have long been buying NetApp filers or EMC Celerras, while SOHO users could take their pick of NAS boxes from companies like NetGear, Buffalo and Iomega. Mid-market organizations have been stuck with the various flavors of Windows including Windows Storage Server or expensive bottom-of-the-line systems from NetApp and EMC. Vendors like OnStor and Reldata have tried to make inroads but haven't been terribly successful, in no small part because customers never wanted to buy block storage and NAS heads from different vendors.
Mid-market block storage vendors like Xiotech and Compellent have offered Windows storage server NAS heads for years, but they haven't been real popular with users. Last year at Compellent's C Drive conference, several users pointedly asked for a faster, more capable NAS solution. Besides limited performance, especially for NFS, the WSS/SAN array combination meant that managing the file server required both Windows and storage skills.
Xiotech went the power route, bundling Symantec's FileStore scale-out NAS solution with their own ISE self-healing array modules. FileStore, based on Symantec's Storage Foundation, uses a cluster-file system to support up to 16 NAS servers in a single cluster. The combination, complete with the now ubiquitous Nehalem processors, provides scale up to 2PB and should have plenty of go-fast for all but the most demanding applications.
Compellent's new zNAS uses NexentaStor, which is based on OpenSolaris, and most importantly, ZFS. This delivers high availability failover, essentially unlimited file system capacity and most significantly, data deduplication.
