Zero Downtime Migration for Casino Administration

Problem:

A large Atlantic City casino used HPE NonStop servers to manage all aspects of the casino’s business, but could never take the application down for any reason due to the 24/7 operational aspects of running a casino. As part of a major upgrade, the casino required developing new applications and the restructuring of its databases. In addition, the NonStop systems needed to be refreshed to newer models. The original databases included both Enscribe files and SQL tables, but the new databases primarily only used SQL tables. The casino needed to create and place into operation the new systems and applications without impacting any of its ongoing operations.

HPE Shadowbase Solution: Zero Downtime Migration

Figure 1 illustrates the process of migrating to a new system, application, and database without any downtime of the existing application. The NonStop server on the left represents the original application/database, and the NonStop server on the right represents the new application/database. Both the old and new applications write data to Enscribe and SQL databases. During the migration, the old and new databases are kept synchronized using Shadowbase data replication. Users are connected to both NonStop systems until the proof and verification phase is complete. Once the new system is tested and verified, users are cut over from the existing system and migrated to the new one, with users experiencing zero application downtime during the system migration. The steps involved in the migration process are described below.

Zero Downtime Migrations

Figure 1 — Shadowbase for Zero Application Downtime via a Zero Downtime Migration

  1. The casino creates the new applications on a separate system and copies the online database to the new system.
  2. The database is kept synchronized with the operational database via data replication.
  3. The new applications are then started and tested thoroughly.
  4. The casino phases over users to the new system without denying service to any of them.
    1. Disconnect users from the old NonStop server
    2. Allow the replication queues to drain (should be very quick)
    3. Connect users to the new NonStop server running the new application and databases
    4. Use Shadowbase to replicate back to the original NonStop server
  5. The old database is kept synchronized with the new database for a period of time for both failover (if a latent issue is detected on the new environment) and capacity needs.

Contact us for more information on this HPE 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.