Homogeneous Active/Active Systems for Regional Bank Service Bureau

Problem:

A large U.S. bank implemented a service bureau in order for its regional banks to process securities and financial information. The service bureau required an active/active architecture that would provide load-balancing and high availability. Due to electrical outages during hurricane season in the Florida node, the bank also required a system that would allow for fail-over of their users to the Midwest node.

collision-avoidance-via-partitioned-users

Shadowbase Solution – Homogeneous Active/Active Systems:

  • Data changes on one HPE NonStop node are applied to the other NonStop node’s database via asynchronous replication.
  • The primary databases for the bank’s customers are assigned to one of the two nodes (Midwest or Florida) in order to load-balance activity across both nodes, which is periodically evaluated and adjusted depending on the volume of the changing transactions.
  • To avoid data collisions using this approach, it is important that the database supports partitioning (via a customer identifier embedded in the record key fields) and that an update for a specific customer is always executed on the same node.
  • Should a node fail, all of its users are reassigned to the second node. By replicating partition changes to the alternate node, a backup access path is provided for updates. This path is only provided in the event of the primary node loss for that partition.

Contact us for more information on this Shadowbase solution

The above was adapted from the book: Breaking the Availability Barrier, Volume III: Active/Active Systems in Practice by Paul J. Holenstein, Dr. Bruce Holenstein, and Dr. Bill Highleyman.