FIXEdge/J Cluster Solution Release Notes

FIXEdge/J Cluster Solution Release 1

LB Cluster solution released on October 1, 2019


EPAM Systems is pleased to announce Load-Balancing Cluster Solution, which improves the scalability of FIXEdge Java - based systems.

It allows to distribute FIX sessions across multiple computing and storage nodes to provide load balanced and reliable processing.

The solution relies on shared state management and API based access to configuration services.

The real-time monitoring of  nodes/applications availability supports uninterruptable service.

Key Features

  • Load-Balancing - input FIX-session flow is uniformly distributed between active FIXEdge Java nodes. 
  • Fault tolerance -  in case one of the FIXEdge Java nodes fails, session flow is redirected automatically to remaining active nodes without messages loss.
  • Centralized configuration management via command-line tool - including add/remove/modify/get list of established sessions commands.
  • Centralized service management based on Consul - to provide component coordination and health checks.
  • Shared FIX storage -  to store FIX messages and all configuration information. Failover Storage Server is used for FIX Session states and outgoing FIX messages.  SQL database (Oracle) is used for configuration data.
  • Centralized logging of FIXEdge Java events – events could be logged to Splunk.

Constraints

Please note, that current solution release is implied to support only acceptor sessions and is available only for Linux OS (Tested on RHEL 7).

Documentation

FIXEdge/J Cluster Solution Release 2

LB Cluster solution released on February 18, 2020

Key Features

The key feature of the current release is the initiator sessions support:

  • Ability to create/update/delete the initiator sessions via CLI Tool in runtime
  • Ability to manage sessions according to a specified schedule (added cluster services component “Scheduler” )
  • Fault tolerance - in case one of the FIXEdge/J nodes fails, initiator sessions are restarted automatically on remaining active nodes without messages loss

Constraints

Please note, that current solution release is available only for Linux OS (Tested on RHEL 7).

Documentation

  • Updated documentation for solution overview.
  • All other documentation links from Release 1 are applicable here.