BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//sf26-agenda//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:SESS-110@sf26-agenda
DTSTAMP:20261020T151709Z
DTSTART:20261105T183000Z
DTEND:20261105T191500Z
SUMMARY:Faster Kafka Without Changing Any Code
LOCATION:Room 2008
DESCRIPTION:Simon Ritter\n\nOne of the primary reasons Kafka is built on Ja
 va is its incredible scalability\, ease of use\, and "write once\, run any
 where" capabilities. However\, the JVM\, a managed runtime environment\, b
 oth helps and hinders in achieving the best possible Kafka performance. As
 pects like garbage collection pauses can make it hard to meet defined SLAs
 . JIT compilation can deliver better overall performance\, but your Kafka 
 nodes take time to warm up. In this session\, we'll explore at a technical
  level how Azul has been addressing these limitations and building a high-
 performance JVM that delivers fantastic results for Kafka. Specifically\, 
 we'll look at: How can GC operate completely concurrently with application
  threads whilst still moving live data within the heap to eliminate fragme
 ntation? What can be done to improve the level of optimisation of code gen
 erated by the JIT compiler Solving the warmup problem. There are several a
 pproaches to this\, and we'll provide a quick comparison of the advantages
  of each and where they can be used effectively Finally\, what if we decou
 ple the JIT compiler and move it to a JIT-as-a-service (or a Cloud Native 
 Compiler)? How can this further improve Kafka applications? The results ca
 n be impressive\, with benchmarks showing over 50% more transactions per s
 econd in certain situations. This session will cover low-level technical d
 etails to help you understand your options when deciding how to optimise t
 he Java environment for your Kafka deployments to maximise performance. By
  attending this session\, your contact information may be shared with the 
 sponsor for relevant follow up for this event only.
URL:https://sf26.rmoff.net/s/faster-kafka-without-changing-any-code-110
END:VEVENT
END:VCALENDAR
