Thomson Reuters Completes a Large-Scale Migration Project Ahead of Deadline on AWS

A few years prior to the divestiture, Thomson Reuters committed to a cloud-first strategy with Amazon Web Services (AWS), taking advantage of Amazon Elastic Compute Cloud (Amazon EC2) and many other AWS services to provide highly secure and scalable cloud-native applications and services to its customers.

However, Thomson Reuters still had not migrated all of its existing applications—many of which were customer-facing—by the time of the divesture, and the Refinitiv data center migration presented an opportunity to accelerate the company’s move to the cloud.

Thomson Reuters decided the best course of action was to make its legacy applications cloud-ready. “It was better strategically to build the infrastructure in AWS and then work with our business units to modernize and redeploy their applications,” says Bart Matzek, senior director of technology in solutions engineering at Thomson Reuters.

The company engaged two managed services providers (MSPs) to assist with the migration, but the project quickly ran into roadblocks. “We were too modest in our own abilities and thought the managed services providers we selected would be able to show us the way and we could learn from them,” says Justin Wright, Thomson Reuters’ vice president of architecture and development. “We sat back and listened as they rolled out their programs and structures with an open mind.”

But Wright and his team soon realized that the migration plans the MSPs came up with were aimed at off-the-shelf software, were too focused on a server-by-server approach, and didn’t include minimum viable refactoring and alignment on common patterns, such as consolidating the versions of operating systems and databases the company was using. “The migration plans from the MSPs seemed to focus on all of the dependencies and things that might go wrong, which sent us down a rabbit hole,” says Wright. “The MSPs’ lack of domain knowledge, coupled with the complexity of our estate, made us realize we needed to put more ownership on the business and its application teams to have them drive their own migration.”

The teams needed to take ownership and reinforce the plan and priorities. "Adding our own discovery process, dubbed ‘permit to host,’ on top of the data that our MSPs collected, allowed us to be very efficient with the architectural patterns we used to provide older database versions and with other technical capabilities our applications relied on," says Matt Dimich, head of solutions architecture at Thomson Reuters. “That enabled us to sort applications into buckets and consolidate patterns of technology, ultimately saving us a lot of time.”

Thomson Reuters engaged the help of AWS Customer Enablement services including AWS Managed Services (AMS) and AWS Professional Services. “We already had a relationship with AWS, and we knew we could rely on its services,” says Wright. “AMS could also provide operational support and resources for capabilities we didn’t have the desire to scale.”

Simon Hornby, senior director of service transformation programs at Thomson Reuters, thought the shift away from the MSP approach was a great opportunity to refocus. “We determined we couldn’t meet the program timelines by refactoring all 400-plus applications, so we needed to consider different migration strategies to continue down the cloud-first path, but without the application transformation. By adopting rehost and replatform migration strategies, we were able to continue with the cloud adoption by making these legacy applications cloud-ready, but more importantly, we were able to meet the primary objective, which was to complete the program within two years.”

One of the biggest challenges with meeting the deadline was that none of Thomson Reuters’ tooling or management infrastructure existed in AWS. “We had no standards for this, as our intention was never to run traditional applications in the cloud. If you look at migrating 10,000 assets and 400 applications, your eyes start to glaze over,” says Danny Stubbings, vice president of engineering and operations at Thomson Reuters. “It’s a daunting amount of work to accomplish in just two years.”

While Thomson Reuters was still responsible for the migration, AWS became the glue that brought the company together and moved the project forward. The AWS Professional Services team first suggested breaking the project down into smaller parts, each with its own dedicated cohort. “This centralized approach assigned responsibility to individual groups and encouraged the cohorts to drive their own initiatives. Each cohort was assigned 5–10 applications to focus on,” says Wright. “The idea was to do lots of small things very quickly.”

Once the cohorts were defined, Thomson Reuters and AWS conducted intensive, clearly focused workshops with experts from various teams, including both AMS and AWS Professional Services. “Those groups helped us push through a lot of roadblocks,” says Stubbings.

Empowering the cohorts was critical to moving the migrations forward. “If we're dragging an application group over the line, it's going to take years,” says Hornby. “But if the application teams are driving the migrations, they will get there on time. Accountability makes a big difference.”

The project required significant collaboration among the teams. AMS worked with Thomson Reuters in an agile fashion to adapt to its operating model while bringing the benefits of change control to iterate and accelerate new capabilities to meet the company’s needs. “We needed a specific feature and it had to happen for us to be successful,” says Matzek. “AMS opened up its infrastructure to us in a special developer mode. We helped shape AMS, which gave us the flexibility to flush out issues quickly, fail fast, make changes, and then smooth out the process. That was a huge win. This continued focus on process improvements and automation helped the pace of migrations increase over time.”

Once the infrastructure was built and refined, Thomson Reuters focused on application migration and the project began to accelerate. In the first five months, Thomson Reuters had migrated just one application. After resetting the overall program structure and collaborating with AWS, the company migrated over 30 applications in two weeks. To further streamline migration, Thomson Reuters used simplified software contracting through the AWS Marketplace to rapidly procure and integrate preferred third-party software into its AWS environment.

Leveraging AWS Customer Enablement services, Thomson Reuters migrated seven globally dispersed data centers, hundreds of applications, and thousands of assets to the AWS Cloud five months ahead of schedule. And because it was able to exit Refinitiv data centers earlier than planned, the company achieved significant savings.
Once this migration was complete, the Thomson Reuters team continued to work with AWS to further modernize the migrated applications, improve operability, and reduce overall cloud costs. This effort resulted in optimized service offerings and additional cost savings of approximately 20 percent. Further progressing these iterative improvements remains a continued focus. The migration to the cloud also increased scalability, improved agility, and helped map a path toward modernization that will position the company to be more innovative. “We're now much better positioned to modernize all our applications,” says Matzek.

Thomson Reuters tackled the daunting project ahead of schedule. “Without the help of AWS, we wouldn’t have met our deadline,” says Wright. “Being able to offload problems to the AWS Managed Services team, and knowing they were a trusted resource that could get us to the endgame, was a huge benefit.”