Legacy business applications are often challenging to migrate to public cloud. Some of the key challenging questions are – Is the application suited for the public cloud? Does the application require any code remediation, re-engineering, re-architecture or can it be migrated as lift-and-shift? How long will the migration take? Can we accelerate the migration process? How do we ensure application on cloud is secure, highly available, fault tolerant, and better performing? Can we realize reduction in TCO once migrated to cloud? Every application is unique but the process, methodology and approach can streamline cloud migration and accelerate journey into cloud.
Dow Jones, a New Corp company is a leader in providing news and information through its prestigious brands including The Wall Street Journal, Dow Jones Newswires, Factiva, Barron’s, MarketWatch and Financial News.
VirtusaPolaris was engaged in application portfolio assessment and migration of Dow Jones applications from on premise into AWS cloud. Our cloud journey included following steps:
Portfolio Discovery and Assessment
Aim of this discovery and assessment was to identify existing IT landscape, identify applications and their owners based on line of businesses. Information was gathered from traditional sources like CMDB, app owners, system administrators etc. We introduced automated discovery tools like, CloudScape, Cloudamize which helped us build a Dependency Structure Matrix (DSM) that provides a compact and clear representation of a complex system and captures method for the interactions between system elements.
VirtusaPolaris has a model to score each application for its fitment to cloud, called cloud suitability index. The score was obtained for each application based on the metrics belonging to three parameters: Cloud Suitability, IT Efficiency, and Business Value.
Based on the assessment, migration workloads were categorized into buckets like Replicate, Rehost, Remediate & Re-engineer. Remaining applications were marked as candidates for consolidation, sunset or retire.
We laid the roadmap for application migration into cloud. Our first set of applications to be moved to AWS were identified. We teamed up with application owners and got the details of the application architecture, system architecture, and gathered baseline performance metrics.
As part of the operation model implementation, we created operation and management plans, identified contacts within the application teams for escalation and support. We collaborated with the application development team to create test plans and user acceptance criteria.
Design and Architecture
Some of the key characteristics and design principles of the AWS architecture we implemented were-
- Dataflow – detailing the connectivity between on-premise and AWS cloud, various services and components.
- Automation – We applied techniques, practices, and tools such as, AWS CloudFormation, AWS OpsWorks, Chef, Jenkins etc. to make the infrastructure reusable, maintainable and extensible.
- High Availability and Disaster Recovery – Each service and components were designed to make it highly available considering the RTO, RPO and cost.
- Scalability – We designed the applications to support growth in users, data traffic, considering the SLAs.
- Databases – Based on the results of the POCs, we designed the applications to consider RDS or use DB on EC2 and laid the roadmap for future enhancements.
- Security – Created security guidelines and best practices. Applications were designed using the proven best practices and real-time auditing and monitoring were implemented.
- Optimization for Cost – Worked closely and collaborated with the application team to benchmark and select the right instance types.
First wave of successful application migration into cloud is critical for laying the path for a factory model.
Server Migration – Our first wave of wins were servers using lift and shift approach, migration of Win 2003 servers into AWS, and applications using migration tools like Racemi.
Application Migration – Applications migrated into AWS had a better latency performance considering the optimally selected instance types and architecture design.
Data Migration – Applications with DB servers like MarkLogic and MS-SQL server were migrated into AWS cloud. Most of the databases were migrated using the ‘lift-and-shift’ move.
Migration Factory Model
In order to accelerate the migration process, we introduced our migration approach that was based on the factory model. Some of the key characteristics of this model were –
- Standardized migration process
- Asset reuse
- Specialized teams
- Scalable core / flex model
- Using automation and tools
We collaborated with the application development team for the application validation, integration validation and business acceptance. Functional testing, Integration testing and performance testing were also done as a part of this process.
Validations were performed on the infrastructure like –
- Security testing (security groups, SSH access, access from corporate network, bastion host access, NAT host access, access to servers on various ports, VPC and subnet level testing on blocked IPs etc.)
- Resilience testing
- Load Balancing
- Web server health checks
- Monitoring and its corresponding notifications
Migration is not complete until the production traffic flows into the instances hosted on AWS cloud. Following considerations were key for the cutover of application migrated to cloud –
- Acceptable downtime by the business
- Any business constraints
- Transactional data; data backups
- Rollback plan
- Change management workflow and approval process
Optimizing the cost is very crucial for applications and services hosted on cloud. Following techniques were adopted for cost optimization as part of our engagement –
- Created scripts to turn off unused instances and services – Used reports from tools like CloudHealth, AWS Trusted Advisor to build scripts for turning off unused instances and services.
- Limited resource provisioning
- Right sized the instances and resources
Last but not the least, program governance, is the key to the success for such a large scale migration engagement. VirtusaPolaris was involved with establishing the governance framework. As part of this engagement, some of the goals of program governance we undertook were –
- Defined and implemented structure involving multiple business units and teams
- Created communication plans and workflow processes
- Provided active direction to the program
- Provided status reporting and regular updates to the program executives
- Did periodical reviews of the interim results for assessment and application migration
- Identified and executed adjustments to ensure achievement of the planned outcome
Results and Benefits
Our solution helped Dow Jones create a well-established process and methodology in identifying applications to be migrated to cloud. We accelerated their cloud journey using migration factory model. Further, we reduced risk by decommissioning some of end-of-life and end-of-support systems as a result of infrastructure refresh as part of cloud migration. Our program governance helped with the seamless migration and cutover. Operational costs were reduced by implementing a disposable and consumption based infrastructure models. Lastly, we reduced the total cost of ownership through enterprise wide solutions, re-usable templates and repeatable process.