Call us: +1-415-738-4000
Garbage collection can trash your application’s performance, causing unpredictable GC pauses that grow longer and more frequent as more heap is used. With BigMemory’s off-heap store, you can provide fast, local access for large amounts of data, all in a single JVM, and all without GC pauses or tuning.
Have you ever spent countless hours tuning GC, only to see your hard work undone by changes in application code, heap size, CPU utilization, or the mix of applications on the server? With BigMemory, the days of iterative, time-intensive tuning are over. In fact, it took one BigMemory customer only three hours to achieve performance results that previously required three months of ongoing tuning.
BigMemory makes it easy to prevent the performance-killing GC pauses that frustrate users and make it hard to meet SLAs. Your applications will run faster and more reliably, with lower, predictable latencies and increased throughput.
In performance tests, BigMemory has demonstrated consistent, “pause-free” application behavior out to hundreds of GBs of in-memory, off-heap cache (see the chart below). By contrast, the same performance tests using an on-heap cache resulted in increasingly longer GC pauses at bigger cache sizes, with full GCs “freezing” the application for minutes at a time.
Many organizations stack lots of 1-4 GB JVMs on a single machine in an effort to minimize the GC problem. This quickly becomes hard to manage and costly to scale.
With BigMemory, you can increase application density—running a smaller number of larger-memory JVMs—and save time and money as a consequence. This simpler deployment model eases application scale out and provides a more sustainable, efficient solution as your dataset inevitably grows.
“Garbage collection has long been Java’s Achilles’ heel… Users will welcome structural solutions to this universal challenge.”— Massimo Pezzini, Vice-president and Gartner Fellow at Gartner, Inc.