In-Depth

IBM Overhauls pSeries

Company introduces disaster recovery features for Unix line

• By Stephen Swoyer
• 10/21/2003

IBM Corp. last week revamped its line of pSeries Unix servers with enhanced performance features and new disaster recovery options.

Chief among the new pSeries enhancements is a Capacity BackUp capability for Big Blue’s high-end 16-way p670 and 32-way p690 pSeries servers. Capacity BackUp systems can be paired with pSeries main systems to provide on-demand, incremental capacity that can be switched on if capacity is saturated or or unavailable, making the systems ideal for disaster recovery scenarios, according to IBM representatives.

The new p670 and p690 Capacity BackUp systems are available with IBM’s “lightswitch” Capacity On/Off Upgrade on Demand feature, which lets customers tap dormant compute capacity to address seasonal or unexpected spikes in demand. The p670 entry ships with four active and 12 inactive Power4+ processors running at 1.45 GHz. The p690 version ships with four active and 28 inactive Power4+ processors, which can range in speed from 1.3 to 1.7 GHz.

IBM says that the Capacity BackUp pSeries systems are available at lower price points—as much as 50 percent lower, in the case of the p690—than the p670 or p690 main systems they’ll be paired with.

Big Blue also announced a new switch designed to enhance pSeries’ credentials as a platform for high-performance computing applications The pSeries High Performance Switch (HPS) is a cluster interconnect that’s designed for use with IBM’s eServer Cluster 1600 environment. It provides greater bandwidth and introduces less latency for Big Blue’s eight-way p655 and 32-way p690 systems. HPS supports up to 16 p655s or p690s.

Speaking of clusters, IBM announced that its Cluster Systems Management (CSM) software—which is used to install, operate, and maintain server clusters—is now available for AIX 5L on pSeries, Linux on pSeries, and Linux on xSeries. CSM provides remote hardware control, utilities for pushing out software updates to servers in a cluster, resource monitoring tools, and supports event triggers which can alert administrators or initiate corrective actions when problems occur.

Finally, Big Blue announced sweeping I/O and RAID enhancements across the pSeries line. All pSeries servers now support Ultra320 SCSI, which supports double the maximum channel bandwidth of the Ultra160 SCSI standard.