Internal IT infrastructure, including an internal data center, was built over many years. The company decided to migrate their applications to the AWS cloud for its scalability, flexibility, and support for automation.
Deployment automation and configuration management was fundamental to the project from its inception, as the company’s leadership correctly identified the typical “lift-and-shift” approach as a dead-end strategy for sustainable operations.
The company’s legacy line-of-business applications have a dependency on MSSQL with MSDTC. An added challenge? MSDTC support and the live migration of a large database to AWS.
The AWS Solution
Per the latest Gartner Magic Quadrant report, AWS has always been a leader in Cloud Infrastructure as a Service with respect to completeness of vision and the highest ability to execute. On top of this, AWS offers multiple services for automation including resources provisioning, configuration management, and application deployments. AWS OpsWorks, AWS CodePipeline, AWS Lambda, and AWS CloudFormation have been used to automate the company’s network monitoring system (Zenoss), application servers, and VPC in the cloud.
The figure below illustrates VPCs in the cloud:
Figure 1: VPCs in AWS with On-premise Network Integration
The figures below illustrate automated deployment of VPC and Zenoss in the cloud:
Figure 2: VPC in AWS
Figure 3: Automated Zenoss Deployment in AWS
Benefits of Automated AWS Infrastructure
Automated deployment of applications brings several benefits both to engineers and managers. When deployment is designed to be repeatable and automated, anyone can perform the deployment, allowing engineers to focus on more critical tasks. Repetition makes the whole process less error-prone, thus benefiting the management implicitly. Moreover, these solutions can be extended to provide for other requirements, such as multi-region deployments or disaster recovery.
The Success Story
By leveraging third-party software, SIOS DataKeeper, the live migration of the production core MS SQL Server database system was made possible. Initially, an EC2 instance was configured as a failover cluster member to the existing production database with volume-level replication across an AWS Direct Connect private interconnect. Failover and failback between EC2-based MSSQL and legacy on-premise MSSQL server infrastructure was exercised. The final outcome was a multi-AZ MS SQL on Windows Server EC2 instances with volume-level replication across AZs and WSFC failover/clustering capabilities including MSDTC.