Presented by MongoDB

---

Financial services companies are often behind the curve when it comes to digital innovation. A combination of regulatory strictures and the complexity of in-house legacy solutions that the business depends on has made IT decision-makers unsurprisingly hesitant to turn to the cloud or serverless architecture.

Midland Credit Management (MCM), one of the largest debt buyer and debt collection companies in the U.S., was in a similar position, mired in relational databases and a highly customized monolithic architecture. They faced an increasingly unstable, unscalable backend environment and solutions, and teams were in constant firefighting mode.

“Because of how complex the environment was, if you wanted to make any changes it was almost impossible to make them without breaking anything,” says Ali Montazer, CTO of MCM.

In a world where digital innovation is increasingly synonymous with competitive advantage, company leaders recognized the need to modernize or fall behind, Montazer adds. It meant changing how the company designs, builds and manages applications which led them to prioritize solutions that enabled their developers to be much more agile.

To meet that challenge, Montazer says, “you need to change the focus from building applications that just serve one purpose and then building on top of that, to more holistically looking at your environment, looking at your use cases, and being able to distribute and scale them as much as possible.”

When tackling their migration and modernization project, MCM examined the three fundamental aspects of a healthy backend architecture: the technology, the people and the processes that support them. They needed to upend their old school notions and completely rethink how the team was organized, how their operating model was actually operating, and finally, how to align the architecture with that organizational change.

Ownership, organization and architecture

Stepping back, MCM pinpointed the issues missing in their organizational model — developer ownership of the features they were building, and a culture of connection and collaboration between the dev team handling implementation and the support team. It meant fostering a sense of accountability for code with their developers, instilling in them a sense of power over decisions. It also meant moving from being entirely project-based to adopting a product mentality — one that developers can own and continue to develop and optimize.

They split the product development team into high-level solution architects, along with an engineering group formed as DevOps/Scrum teams, led by a Software Development Manager (SDM), a Technical Product Manager, a Scrum Master, and developers and testers. The scrum teams act like startups inside the enterprise, responsible and accountable for the products they build, with ownership over the whole software development lifecycle.

With the organizational changes settled, MCM turned their eyes toward drastically simplifying the architecture supporting the development, launch and support for applications.

The goal was to move from an on-premises monolithic environment to an event-based distributed cloud-native operating model. The new microservices model would enable developers to build applications utilizing serverless technology for large amounts of streaming data.

The inaugural project would be creating a single-view application which would unify disparate data sources for omnichannel campaigns, such as regulatory mandated customer communication about their debts.

Modernizing with MongoDB and AWS

Many of the company’s financial systems and services were in-house solutions, which have on-prem hosting requirements. It wasn’t possible to do a lift-and-shift migration into the cloud, and it wasn’t scalable.

AWS and MongoDB made it possible for MCM to continue using their own on-prem database by pushing database modifications through CDC (capturing data changes) and streaming it with Amazon Kinesis to the AWS cloud. MongoDB Atlas receives the data and acts as an aggregator to ensure every set of data sent and received is in the right state.

Every data shard has a timestamp; MongoDB puts those together in a document, which is then ready for the events that MongoDB generates for the microservices that process the data. Rather than putting everything in one document, which can grow expensive if the number of data fields keeps expanding, the process can be modeled like an API call, where the data is sent as needed to generate an event.

The real-world result of MongoDB and AWS implementation

One of MCM’s legacy solutions is a service that generates and sends welcome and validation letters to consumers. Due to regulatory requirements, these mailings need to go out within a certain timeframe.

Before they implemented their new solution, MCM was only able to generate 14,000 letters in a single run. It would take up to five days to process one million letters, and 450 executions per month to meet the required volume.

With the MongoDB and AWS serverless solution in place, each run generates 720K letters — a 50x improvement in scalability along with being able to finish the task 5x faster than before.

They were also able to add a rule engine to the service, which means that rather than having to go to a developer to ask for a code change when any updates need to be made, they can use the rule engine. Finally, the bandwidth of the whole solution has become completely distributed, so that they’re able to scale up as needed for the few days in the month when they send letters, and then scale back down, significantly reducing the impact on the company’s main service.

“Being able to leverage these new solutions built on MongoDB Atlas on AWS has allowed our development teams to become much more productive and utilize modern software development practices,” said Montazer. “At the end of the day, we’re able to deliver better experiences to our end-users through the new software solutions we’ve brought to market.”

Learn more here about MongoDB Atlas, their multi-cloud developer data platform, and try it free forever here.

---

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact [email protected].