Consistent, Fast Response Times

When critical business applications pause, companies lose money. When it comes to fulfilling on-line purchases, executing stock trades at the real time price, acting on price fluctuations or approving loan applications, completing only 85 percent of the requests in time is a failure. When an application doesn't respond in a timely manner, the customer goes elsewhere and may not come back. Revenue is being lost. Yet many businesses today are living with 15 percent or worse incompletion rates due to application pauses caused by scalability or memory limitations. It's time to stop leaving money on the table.

For a wide range of business critical enterprise Java applications, Azul Compute Appliances has dramatically reduced the worst-case response times as compared to traditional servers, all while enabling applications to scale to significantly higher levels of throughput. With Azul Compute Appliances, worst-case response times are typically reduced by a factor of 3x to 10x as compared to UNIX and x86 servers, enabling applications to deliver much more consistent performance even for highly unpredictable and spiky workloads.

	Application response time is key to closing the most business in the shortest amount of time. On traditional servers, shown in Orange, nearly every application reaches a saturation point – the point where application response time is at its worst. Putting these applications on Azul eliminates processing and memory scalability bottlenecks and dramatically lower s peak response times. In these four real world customer examples, deploying on Azul lowered peak response times by as much as 10 times.

The reason many high-volume Java applications do not consistently meet customer SLAs is because of memory management limitations of Java Virtual Machines (JVM). Periodically, when the JVM memory gets full, the old data must be cleaned out to make room for new content in a cleaning process known as garbage collection (GC). On traditional servers, this process requires the application to pause and can lead to minutes-long pauses if the JVM memory is larger than the practical limit of 2 GB.

Azul Compute Appliances eliminate GC-related application pauses and allow single instances to tap into memory footprints as large as 670 GB . This is made possible by the industry's only hardware assisted Pauseless Garbage Collection, a series of GC algorithms that rely on unique instruction sets in the Vega 3 processor and allow the JVM memory to be cleaned without pausing the application.