IBM swims downstream with 'baby Shark'

By Heidi Biggar

At the Storage Networking World conference and exhibition last week, IBM fleshed out its enterprise-class disk system family with the announcement of a lower-cost version of its TotalStorage Enterprise Storage Server (ESS) 800 "Shark" array.

Dubbed "baby Shark," the ESS 750 has many of the enterprise-class features of the ESS 800 array but is scaled to meet the price, performance, and capacity requirements of mid-tier users. The starting price for the new system, which scales from 1.1TB to 4.6TB, is $100,000.

Like its predecessor, the ESS 750 can be used in an open-systems storage area network (SAN) environment or in mainframe environments. (It can also be attached to a network-attached storage [NAS] gateway to serve both block- and file-level data.) In contrast, IBM's FAStT arrays work only in open-systems environments.

The Evaluator Group consulting firm divides storage systems into two categories: those that are cache-centric (i.e., very large centralized caches) and those that are distributed (i.e., can only be attached to open-systems servers).

According to the Evaluator Group, examples of cache-centric systems include EMC's Symmetrix DMX, Hewlett-Packard's XP, Hitachi Data Systems' 9900 V, IBM's ESS, and StorageTek's SVA. Examples of distributed systems include EMC's CLARiiON, Hitachi's 9500, HP's EVA, IBM's FAStT, StorageTek's D-Series, and Sun's 6320 array.

As for how these two general category of products compare, Chuck Standerfer, senior partner at the Evaluator Group, says that cache-centric systems are generally more complex and higher priced (due to mainframe-specific features and functions) than distributed systems, although distributed systems arrays are often just as reliable, fast, and feature-rich as their cache-centric counterparts. The ESS 750 has 8GB of cache.

IBM is positioning the new ESS 750 array for two types of customers: those that already have mainframe storage and need to add capacity, and mid-tier open-systems users that are price-sensitive and may be looking for advanced business continuity features such as replication.

With the ESS 750, IBM says it has filled a hole in its ESS lineup and has jumped ahead of the competition in at least one area--scalability, claims Brian Truskowski, general manager of IBM TotalStorage Open Software.

Unlike some competitors' platforms, users can field-upgrade to a higher-level platform (in this case, from the ESS 750 to the ESS 800), according to Truskowski.

For users looking to scale incrementally, IBM offers Standby Capacity on Demand (Standby CoD), which allows users--particularly those in unpredictable environments--to pay a nominal fee up-front to keep extra capacity on tap. Users activate the standby capacity when needed--without IBM intervention--and are charged for the capacity once it has been logically configured.

Highlights of the ESS 750 include support for FlashCopy, Peer-to-Peer Remote Copy (PPRC) and, when deployed in zSeries environments, Parallel Access Volumes, Multiple Allegiance, and Priority Queuing for more-efficient data movement and usage. The ESS 750 does not support Extended Remote Copy (XRC).

Other hardware features include a two-way processor, RAID-5 and RAID-10 support, and up to six Fibre Channel/FICON or ESCON host adapters.

This article was originally published on April 14, 2004