Migrating Workloads to AWS for leading online bookmaker

Migrating Workloads to AWS for leading online bookmaker

The opportunity at a glance

Driving risk reduction, improving efficiency and providing capability uplift.

Industry

  • Wagering / Gaming

Customer overview

Our customer is Australia’s biggest online corporate bookmaker. 

Business challenge

Gaming is a highly competitive industry. To enable our customer to obtain a competitive advantage in the market, and to iterate upon products quicker to attract and keep punters using their platform, the business has adopted an all-in approach to Cloud.

The organisation had been augmenting its environment into AWS in a very ad hoc, bespoke and not overly secure manner. They committed to an all-in cloud migration which included:

  1. Building new AWS infrastructure in an automated repeatable and secure fashion
  2. Migrate applications from legacy AWS accounts into the new infrastructure using the new patterns designed
  3. Migrate applications from on-prem into new infrastructure using the new patterns designed.


These patterns aim to: 

  1. Improve the security posture of the infrastructure in AWS
  2. Reduce the running costs of the infrastructure by implementing appropriate scaling 
  3. Improve business agility in highly competitive business segment.

Solution

Cevo’s skills and capability within AWS were engaged to help our customer migrate their workloads, specifically:
  • Cloud skills and capacity uplift within the delivery teams
  • Develop patterns for build, package, deployment, logging, monitoring and alerting for new and existing applications in AWS
  • Migration of existing workloads into AWS using the defined patterns
  • Assist with new builds into AWS using the defined patterns 
The framework for migration of applications consisted of four key phases that needed to be undertaken and completed:
PHASE 1
  • Baseline existing on-prem performance and health by introducing Introduce application performance monitoring into the op-prem solution 
PHASE 2
  • Design, build and implement the appropriate solution for the application within AWS. This consists of:
    • Using AWS services (for instance SQLSERVER RDS instead of a self-managed SQLSERVER) when appropriate
    • Minor application refactor to support immutable ephemeral solutions, particularly around how application configuration is managed
    • Implement secure a password manager solution for database passwords
    • Introduce ‘observability’ by adding tools to be able to diagnose issues without needing to ‘log onto the box’
    • Minor to major application uplift depending on the business criticality and longevity of the application
PHASE 3
  • Cut over to the new solution
PHASE 4
  • Decommission the old solution

During the migration process, Cevo implemented changes to the application toolchain including:

  • Introduction of immutable and ephemeral application servers away from domain joined Windows hosts
  • Changed the Continuous Integration and Continuous Delivery solutions of Bamboo & Octopus deploy to Jenkins with native CloudFormation.
  • Changed the database technology from InfomixDB to Aurora MySQL RDS
  • Implemented ‘observability’ application performance monitoring using NewRelic APM, NewRelic Synthetics and SumoLogic
  • Implemented business SLAs via CloudWatch, NewRelic and PagerDuty

Benefits

This project allowed our customer to realise a number of significant benefits around efficiencies, risk reduction and capability uplift:

  • Prior to this migration, manual intervention was required to manually scale infrastructure relative to key gaming events – now they can leverage AutoScaling to scale upon demand
  • Scaling and recovery times for an application server has gone from 30 minutes down to 5 minutes
  • Changes to application code can be deployed via Jenkins all-the-way to production within 30 minutes
  • The trust in the running application is greatly improved by using the exact same application code, version and topology throughout all environments. If there is a production problem, a first option is to kill that instance knowing another will immediately start up
  • With the introduction of Observability, they now have a deep understanding of the application performance of the solution
  • The solution has been designed to be easily replicable for other sports that have the same betting options.
Read more case studies