If Sleepycat Software CEO Mike Olson has his way, the company's Berkeley DB, an open source developer database,...
will find its way into nearly every telecommunications switch and router, and mobile and online application on the market.
Berkeley DB, an embedded database, is well known in the telecommunications industry and many Web sites, where the software acts as the core message store for e-mail, messaging and Web security identification applications. Sleepycat also makes Java and XML editions of the database used by Amazon.com, Google, Cisco Systems Inc. and British Telecom.
Berkeley DB came out of research at the University of California, Berkeley, in 1990. Sleepycat was formed in 1996 because there was commercial demand for an improved version of the academic project, Olson said.
"A number of companies needed to produce directory software and it was recognized as a good product but not enterprise grade," Olson said. "Today there is an enormous amount of innovation, especially in mobile telephony."
The Lincoln, Mass.-based company this month is rolling out the next version of Berkeley DB, version 4.3, which adds support for in-memory transaction logging and automatic sequence number generation. Both features aimed speeding data in the telecommunications industry.
The company also recently opened offices in London where Olson said salespeople are beginning to move aggressively for customers in the booming mobile communications industry in Europe.
"Much of the innovation we're seeing with mobile telephony is in Europe," Olson said. "Text messaging and multimedia messaging there is huge and the infrastructure there is allowing for some very successful deployments."
Sleepycat is seeing a resurgence from its early days in the 1990s. A renewed interest in open source databases and a surge in commercial products being open sourced placed the spotlight on Berkeley DB, Postgres and other small DBMSes.
Sleepycat faces virtually no competition in the space, according to industry analysts. IBM's recently open sourced Cloudscape Java DBMS comes closest to having similar capabilities to Berkeley DB, but Sleepycat has found its niche, said Noel Yuhanna, an industry analyst with Cambridge, Mass.-based Forrester Research Inc.
"This is not your typical database," Yuhanna said. "This small embedded database plays an important role and it has an edge in the open source world."
Today the embedded database is leaving its mark quietly on the Internet, helping applications track users names, passwords, preferences, and personalization data. In addition to providing user authentication and permissions to track large data volumes, Berkeley DB also manages data to link key words, links, relationships, structure and textual and non-textual data.
Berkeley DB boasts that it can support thousands of simultaneous threads of control language that manipulates large databases. Its simple architecture still supports features common in larger DBMSes, such as replication, hot backups and fine-grained locking, Yuhanna said.
The latest version adds in-memory transaction support, which logs transactions in memory to reduce latency, Olson said. Automatic sequence number generation, a feature that has finally been added, is available in most SQL engines and also speeds processes.
Despite being open source, Berkeley DB has a team of developers devoted to adding new features, Olson said. It is available with source code under a modified BSD-style software license, which, once downloaded, can be used, copied, studied, modified and redistributed. Developers who want to redistribute DB with proprietary applications must license it from Sleepycat.
Olson said he is not surprised by the commoditization in the database market. Proprietary database software vendors are having trouble sustaining the high margins from their legacy licenses since the market is saturated, Olson said.
"IBM is very sophisticated about open source and there's no surprise that they have released the Cloudscape code because it was not a huge revenue generator for them," Olson said. "I'm pleased to see more products get open sourced because the sharing of source code gives more insight into how these products work and it forces developers to create better products."