By Lisa Coleman
Hewlett-Packard is increasing performance and cutting prices on its next generation of StorageWorks network-attached storage (NAS) devices. HP also introduced multi-pathing software to improve storage area network (SAN) connectivity and NAS manageability.
The StorageWorks Version 2 products—b2000 v2, b3000 v2, and e7000 v2—all use 2.8GHz Intel Xeon processors and are based on HP's ProLiant DL380 G3 family of servers, which HP claims boosts performance by about 20%. The NAS servers also have snapshot technology based on Columbia Data Products' Persistent Storage Manager software.
Prices on all three Version 2 devices have been lowered. The price of the b2000, a SCSI-based 2U system with three 146GB drives, was reduced from $12,000 to $7,995; the b3000 (which connects to the StorageWorks Modular SAN Array [MSA] 1000), from $44,995 to $24,995; and the high-end e7000 (a NAS head that attaches to HP disk arrays), from $48,000 to $34,995.
The price cuts are due primarily to integration of industry-standard components, according to Mark Nagaitis, director of product marketing in HP's infrastructure and NAS division. All of the servers are based on Microsoft Windows-powered NAS software. HP and other OEMs will move to the new Microsoft Server 2003 operating system (released this month) within three to four months.
HP also added NAS Data Path Manager software to its Linux-based StorageWorks NAS 8000 server to improve connectivity into SAN environments by allowing data-path fail-over between host bus adapters (HBAs) and by providing load balancing and performance monitoring. The user interface allows routes to be designated as active or fail-over paths, and the software also reports failures and displays path status.
Separately, HP plans to migrate Lustre technology, an open-source file system, onto the NAS 8000 next year. Lustre is a Linux-based high-performance, scalable file system being developed by Cluster File Systems, in Mountain View, CA. Lustre is a type of object-based storage, which analysts say can reduce bottlenecks because servers can find applications without searching through the file system.