Announcement Announcement Module
No announcement yet.
Integration performance and memory allocation Page Title Module
Move Remove Collapse
Conversation Detail Module
  • Filter
  • Time
  • Show
Clear All
new posts

  • Integration performance and memory allocation

    Hi all,

    I'm building what is, in effect, an extremely high throughput, low memory allocating (goal of 0 or v.infrequent full GC), near real-time (microsecond latency front to back) message processing system which is made up of various apps which perform similar jobs, but are specialised for a particular function. These apps are in turn built up of similar core components which will be assembled and configured for their particular function.

    The assembly could get fairly involved, with adapters on each end of a component and potentially needing the option to wire up components in local or remote configuration. i.e. be able to route the results of one component to the input of another local/remote service, with remote comms being handled by an external message bus.

    In other words, a good fit for Spring Integration!
    • Does anyone have experience of using the framework in this type of application?
    • Post start up, how involved is the framework in message processing? i.e. what are the performance implications (throughput & latency) of using the framework for messaging?
    • Considering we will be using an external message bus, is Integration overkill and would we be better simply using Spring and implementing component wiring ourselves?

    Many thanks for any help in advance