Bank of Georgia Logo
Industrie : Banking & Financial Services
Lieu : Georgia
Télécharger le témoignage entier

The Bank of Georgia transforms its infrastructure to deliver best-in-class customer experiences

Points clés

  • Four to five times faster processing times.
  • Zero failures during periods of peak demand.
  • Systems scale automatically based on demand.
  • Single pane-of-glass centralizes management of complex environment.
  • Automated and centralized logging system simplifies the debugging process.

Produits

Introducing the Bank of Georgia

The Bank of Georgia is one of Georgia’s largest companies, providing its banking and financial services to more than half of the Georgian population. Listed among the 250 largest and most successful companies in the London Stock Exchange, the company manages 21.8 billion lari in assets ($7.1 billion USD), employs over 6,200 people and is known for its digital banking services. Credit for its popular mobile and digital applications belongs to a fully in-house development team of 700 IT professionals who also built the company’s core banking system, digital channels, payments channels and more. “We began in 2017 to change the way the company thought about itself,” says Vazha Mantua, deputy CIO at Bank of Georgia. “We’re not just a bank; we’re a technology company.”

At-a-Glance

The Bank of Georgia’s commitment to delivering best-of-breed solutions to its customers instigated a decision to transform its monolithic processes and applications into portable microservices. Aiming to hasten development velocity and to be cloud-ready within five years, the bank sought for a Kubernetes management solution that could support its complex cluster environment. The company chose Rancher Prime to manage its Kubernetes estate. As a result, the bank is seeing 5x faster processing times, as well as increased efficiency, scalability and security, hastening the development velocity of new and innovative services.

Transforming into a digital bank

When Bank of Georgia built its IT infrastructure in the early 2000s, it was made to support bank operators working standard banking hours. But as demand for online and mobile banking services grew throughout the 2010s, the IT team retrofitted its environment to deliver on the bank's outstanding customer experience.

Maintaining this infrastructure, however, was time consuming and expensive, demanding untold people hours for applying manual patches and updates. This threatened to not only stymy the company’s ability to develop innovative services, but also interrupt a continuity of service expected in the digital age. As a business committed to innovation, Bank of Georgia knew its next challenge would be to become more agile and that its technology infrastructure must scale to hasten development velocity.

“We had a lot of failures during periods of peak demand,” describes Mantua, referring to holiday shopping seasons. “Our core banking system wasn’t built to support spikes in usage from our mobile and digital channels. We needed to increase our scalability in order to meet this demand.”

The team also wanted to simplify management of its security parameters to defend against ever-evolving online threats and protect customers’ private information.

“As a bank, we are responsible for keeping our customers’ accounts and information secure. We, therefore, have very strict security requirements,” says Vazha Pirtskhalaishvili, head of the engineering unit at Bank of Georgia. “We wanted to manage all security details from one place, whether it be network policies, accessibility or simply safeguarding our services.”

To address these challenges faster, the company embarked on a journey to transform itself from a traditional banking company, supported by monolithic IT processes, into a digital banking company, supported by DevOps methodologies and microservices.

“We run a very large IT environment — we have more than 2 million lines of code in our core banking system,” says Mantua. “So, we decided to split our core banking system into microservices to speed time to launch and deliver new innovative offerings.”

"With Kubernetes and Rancher Prime, we get the scalability, mobility and high availability we need to deliver high-performing solutions to our customers."

Adopting Kubernetes and Rancher Prime

The team turned to its IT consulting partner of seven years, High-Tech Solutions (HT Solutions), to help build a Kubernetes infrastructure and source a Kubernetes management solution that would fit the bank’s requirements and deliver on four key needs: high availability, performance, scalability and mobility.

After researching potential Kubernetes management solutions over several months, the company asked for a proof-of-concept (POC) from top three contenders, Rancher Prime being one of them. The POC lasted for more than three months as the team thoroughly tested any possibility to ensure optimal performance. As the POC ended in June 2021, the team selected Rancher Prime.

“The main reason why we chose Rancher Prime is because it is upstream Kubernetes with additional features, and its enterprise support pricing model works for us,” says Mantua.

Rancher Prime’s one-node licensing structure enables IT departments to scale, whether it be on-premises or in the cloud, without incurring any additional costs. This frees the team to innovate and experiment without financial consequence.

“Some vendors add too many features to their flavor,” adds Pirtskhalaishvili. “Rancher Prime has the right features that we need, is more flexible and offers additional tools that make our lives easier.”

By September the same year, the team completed building its Rancher Prime and Kubernetes environment. In the following months, it built two clusters (one for development and the other for production) and completed migrating its core banking applications to the new infrastructure. Impressively, the team also implemented a stretched Kubernetes cluster over two data centers that sit miles apart — this stretched cluster is the first of its kind in the region.

“Kubernetes and Rancher Prime provide the kind of flexibility we need for managing a complex, hybrid environment running portable microservices across our two data centers,” says Mantua.

Pirtskhalaishvili credits reaching these milestones in record time to colleagues at HT Solutions: “The High-Tech Solutions’ engineering team had experience with large Kubernetes deployments and Rancher Prime in the past. With this team by our side, we didn’t need to do a lot of research in advance, so this sped up the process quite a bit.”

A cloud-ready future

After implementing Kubernetes and Rancher Prime, the bank’s infrastructure is faster, more secure and runs with minimal intervention from its staff. “Before, we would typically spend a day on administrative tasks,” says Mantua, “But now, our administrative tasks take only a couple of hours to complete.”

Speed and flexibility

By setting up a multicluster, microservices-based environment running in Rancher Prime, the bank’s DevOps team can separate and manage its applications in isolation but still have them side-by-side within a central management platform.

The environment also enables the team to develop new applications as standalone cloud-native microservices that can run in most environments or any cloud. In other words, applications are now portable and can scale in an instant; they can also be managed and updated away from the infrastructure core, improving stability.

The bank’s systems also process four to five times faster, and the team now has the flexibility to innovate and apply updates, or tear down services, seamlessly.

“When we have a new version of an application, we want to get it into our customers’ hands as quickly, as securely and as error-free as possible,” says Pirtskhalaishvili. “Kubernetes and Rancher Prime allow us to apply these updates in such a way that the customer doesn’t notice, and that’s very important.”

Increased efficiency and autoscaling

Equipped with Rancher Prime, the team is now able to automate several basic processes so that developers can spend less time on administration and more time on application innovation.

“It is important for us to have our critical systems increase or decrease capacity automatically for productivity,” says Pirtskhalaishvili. “Kubernetes allows us to serve dynamic customer flows seamlessly and with optimal resources. It also makes our IT infrastructure smart enough to eliminate problems (self-healing) or scale to meet the needs of the business without human intervention.”

Initially, team members dreaded the project ahead due to the amount of work involved, but when Pirtskhalaishvili and Mantua told them about Rancher Prime’s autoscaling abilities, the team members got excited, speaking to how important the feature is to the team.

“We utilize fewer resources now,” continues Pirtskhalaishvili. “In the past, we had a lot of idling VMs because we wanted to be ready for more demand. Now, we have it set up in such a way that if demand fluctuates, it upscales or downscales automatically, depending on need. It’s very flexible and fast, and it has freed up our resources a great deal.”

Security and visibility

By removing this administrative layer just described, Rancher Prime brings developers closer to the architecture itself and deeper into a cluster’s performance. Hence, automating access controls and rules-based security policies mean less time configuring and more developing.

To set these rules properly, project leads worked with the bank’s IT security team to embed security policies into the architecture.

“When security is built into the application layer in Kubernetes, you get more visibility and functionality,” says Pirtskhalaishvili. “This gives us and the IT security team more visibility into configurations, restrictions, which services tap into other services, and costs incurred.”

Mantua adds: “Kubernetes also prevents unnecessary horizontal traffic, which we have. It’s a very big benefit for us.”

The same single pane-of-glass used to view and manage security settings also contains a centralized logging system to make fixing issues easier.

“The development team also likes the centralized logging system because they have everything they need to debug a problem in a centralized location,” says Mantua. “This helps them find issues and debug problems faster.”

What’s next for Bank of Georgia?

Despite its rapid progress to date, a lot of work remains ahead before the Bank of Georgia is cloud-ready.

“We’ve separated our applications into smaller microservices, but there’s still room for them to be broken down even more before they’re cloud ready,” says Pirtskhalaishvili. “Our microservices will be more like mobile components we can put anywhere we want, and they’ll run perfectly. In this regard, the biggest benefits for us are still ahead.”

The Bank of Georgia is excited for its digital future. In addition to finalizing its preparations for migrating to the cloud, the IT team hopes to recruit a slew of IT talent from the region who can help manifest the bank’s digital ambitions and beyond.