The most popular and comprehensive Open Source ECM platform
Let us know more
Technology: Rev up your Java Engines
Since Formtek | Orion’s Java-based engine is at the heart of our Orion 5 technology,
we’re always interested in developments happening on the Java platform, like differences
in Oracle, JBoss, WebLogic, and IBM’s WebSphere application servers. But now there’s Java with a new twist — a company selling hardware and software tuned for large-scale Java applications.
Azul Systems is targeting resource-hungry Enterprise Java applications head-on with their
line of multi-processor multi-core single-box appliances that run a tuned version of the Java Virtual Machine (JVM). They’ve just announced the third generation of their specialized grid-architecture hardware.
Their top-of-the-line model 7280 has 16 chips for a whopping total of 768 processor cores, 768 Gigabytes memory and consumes 3235 watts. That’s roughly equivalent to 33 dual-chip, dual-core x86 servers. The Azul model 7240 has half the capacity and consumption of the 7280. Both of these new appliances run on a special 48-core Vega 2 chip the was especially
designed by Azul.
Azul claims that because their appliances use a tuned proprietary virtual machine proxy technology and their own chipset they can deliver 5x to 50x scale and throughput improvements over general-purpose servers. Further, the appliances can increase the amount of cached data by a factor of 50.
The Azul hardware runs a fully compliant JVM called the Azul Virtual Machine (AVM).
This compatibility allows any application to move to the Azul system without having to
change the application code.
Applications running on a standard JVM begin to reach limits when scaled to about
2 Gigabytes of memory with 4 to 8 processors. Much of the problem is related to garbage collection trying to deal with the large amounts of memory. Azul’s technology boasts of having solved this problem and offer what they cause ‘pauseless garbage collection’.
Azul technology brings predictability, consistency, and scalability to Java-based solutions.
It’s another example of how processing power is growing dramatically while computing costs are simultaneously dropping.
