GEMSTONE’S GEMCACHE DATA TRANSACTIONS SOLUTION DEMONSTRATES SUB-MILLISECOND PERFORMANCE IN ZERO-LOSS TRANSACTION BENCHMARK

Proven Benefit to Financial Institutions Seeking Ultra Low-Latency for Non-Stop Trading Environments

Beaverton, Ore. – January 14, 2008 – GemStone Systems, Inc., the leading provider of the Enterprise Data Fabric™ (EDF™), today announced the availability of the GemCache data transaction benchmark with Intel and IBM.  The benchmark, conducted by the Securities Technology Analysis Center (STAC®), tested the distributed in-memory caching solution in a transactions environment measuring throughput and latency under varying rates of order traffic.

GemCache Data Transactions is designed for high performance transaction environments such as electronic trading.  In the single-server test, the GemCache solution was configured with GemFire™ Enterprise 5.0 software running on an IBM BladeCenter® HS21 server with dual Quad-Core Intel™ Xeon™ 5355 processors. GemCache demonstrated real zero-loss reliability and guaranteed messaging into the "transaction", exceeding traditional criteria set for transaction-only benchmark tests.  In the multi-server tests, a Cisco InfiniBand switch, integrated in the IBM BladeCenter® was used to provide a high performance interconnect.  The project included tests of several GemCache configurations with and without replication at transaction rates varying from 100-6000 orders per second. 

The benchmark results showed the best-of-breed solution was able to support latency of less than 1 millisecond (987 msec) at 6000 orders per second, with full data replication using a two-server GemCache configuration.  In the tests, GemCache was also able to deliver a mean latency of 519 microseconds at 6000 orders per second with a single server without replication.  Additionally, three active/active two-server clusters supported 18,000 orders per second with approximately 1 millisecond (1017 msec) of end-to-end latency using six IBM blade servers.

“The combination of automated trading and regulatory changes on both sides of the Atlantic has brought a rapid increase in the volume of transactions and transaction-related messages, which makes it much harder to keep latency in check,” said Peter Lankford, founder and director of STAC.  “The kind of sub-millisecond performance that GemCache has achieved will be important as trading firms strive to contain or shrink execution latency in an increasing number of markets.”

Through this strategic collaboration, GemStone Systems, Intel, IBM and Cisco are able to address common issues of latency, resiliency, and scalability in every asset class and geography.  Rather than using a central relational database, the GemCache solution is based on distributed in-memory caches, providing seamless horizontal scale-out and transactional throughput. 

“Trading firms today are searching for more efficient approaches to key workloads and eliminating lost orders,” said Juan Menendez, senior vice president, Business Development and Alliances, GemStone Systems.  “Most trading systems are based on distributed architectures coupled by messaging; however, GemStone proposes an alternative persistence model based on distributed in-memory caches rather than a central relational database.  As a result, the GemCache solution will drastically accelerate trading while maintaining zero data loss and high data availability while ensuring complete fault tolerance between data centers and increasing revenues in mission critical areas.”

With direct market access to the exchange, financial institutions require ultra-low-latency data infrastructures that can instantly distribute required information to all of its trading platform systems.  As trade volumes continue to increase, the benchmark proves that GemCache is capable of providing an extremely low level of latency critical for trade executions at favorable price points and achieving a zero failure environment for non-stop trading.