Financial services use cases for IBM Cloud

Context

To improve risk analysis, the company looked to IBM Cloud Kubernetes Service and IBM Cloud Analytic services to reduce costs, increase worldwide availability, and ultimately accelerate regulatory compliance. With IBM Cloud Kubernetes Service in multiple regions, their analysis apps can be containerized and deployed across the globe, improving availability and addressing local regulations. Those deployments are accelerated with familiar open source tools, already part of IBM Cloud Kubernetes Service.

They started by containerizing the analysis apps and putting them in the cloud. In a flash, their hardware headaches went away. They were able to easily design Kubernetes clusters to fit their high-performance CPU, RAM, storage, and security needs. And when their analysis apps change, they can add or shrink compute without huge hardware investments. With the IBM Cloud Kubernetes Service horizontal scaling, their apps scale with the growing number of records, resulting in faster regulatory reports. IBM Cloud Kubernetes Service provides elastic compute resources around the world that are secure and capable.

Now those apps receive high-volume data from a data warehouse on IBM Cloudant. Cloud-based storage in IBM Cloudant ensures higher availability than when it was locked in an on-premises system. Since availability is essential, the apps are deployed across global data centers: for DR and for latency too.

They also accelerated their risk analysis and compliance. Their predictive and risk analytics functions, such as Monte Carlo calculations, are now constantly updated through iterative agile deployments. Container orchestration is handled by a managed Kubernetes so that operations costs are reduced too. Ultimately risk analysis for mortgages is more responsive to the fast-paced changes in the market.

• Heightened need for better financial risk management drives increases in regulatory oversight. The same needs drive the associated review in risk assessment processes and disclosure of more granular, integrated, and abundant regulatory reporting.
• High Performance Computing Grids are the key infrastructure components for financial modeling.

The problem of the company is scale and time to delivery.

Their current environment is 7+ years old, on-premises, and with limited compute, storage, and I/O capacity. Server refreshes are costly and take a long time to complete. Software and app updates follow an informal process and aren't repeatable. The actual HPC grid is hard to program against. The API is too complex for new Developers who join the team and requires non-documented knowledge. And major app upgrades take 6 - 9 months to complete.

Solution

Compute, storage, and I/O services that run in public cloud with secure access to on-premises enterprise assets as needed.

• Secure and scalable document storage that supports structured and unstructured document query
• Lift and shift existing enterprise assets and app while they enabled the integration to some on-premises systems that won't be migrated
• Shorten time-to-deploy solutions and implement standard DevOps and monitoring processes to address bugs that affected reporting accuracy

Technical solution:

• IBM Cloud Kubernetes Service
• IBM Cloud Object Storage
• IBM Cloud Data Engine (Spark)
• IBM Cloudant
• Secure Gateway

IBM Cloud Kubernetes Service provides scalable compute resources and the associated DevOps dashboards to create, scale, and tear down apps and services as needed. Using industry-standard containers, apps can initially be re-hosted on IBM Cloud Kubernetes Service quickly without major architectural changes.

This solution provides the immediate benefit of scalability. By using Kubernetes's rich set of deployment and runtime objects, the mortgage company monitors and manages the upgrades to apps reliably. They're also able to replicate and scale the apps that use defined rules and the automated Kubernetes orchestrator.

Secure Gateway is used to create a secure pipeline to on-premises databases and documents for apps that are re-hosted to run in IBM Cloud Kubernetes Service.

IBM Cloud Object Storage is for all raw document and data storage as they go forward. For Monte Carlo simulations, a workflow pipeline is put in place where simulation data is in structured files that are stored in IBM Cloud Object Storage. A trigger to start the simulation scales compute services in IBM Cloud Kubernetes Service to split the data of the files into N event buckets for simulation processing. IBM Cloud Kubernetes Service automatically scales to N associated service executions and writes intermediate results to IBM Cloud Object Storage. Those results are processed by another set of the IBM Cloud Kubernetes Service compute services to produce the final results.

IBM Cloudant is a modern NoSQL database that is useful for many data-driven use cases: from key-value to complex document-oriented data storage and query. To manage the growing set of regulatory and management report rules, the mortgage company uses IBM Cloudant to store documents that are associated with raw regulatory data the come into the firm. Compute processes on IBM Cloud Kubernetes Service are triggered to compile, process, and publish the data in various reporting formats. Intermediate results common across reports are stored as IBM Cloudant documents so template-driven processes can be used to produce the necessary reports.

Results

• Complex financial simulations are completed in 25% of the time than was previously possible with the existing on-premises systems.
• Time to deployment improved from the previous 6 - 9 months to 1 - 3 weeks on average. This improvement occurs because IBM Cloud Kubernetes Service allows for a disciplined, controlled process for ramping up app containers and replacing them with newer versions. Reporting bugs can be fixed quickly, addressing issues, such as accuracy.
• Regulatory reporting costs were reduced with a consistent, scalable set of storage and compute services that IBM Cloud Kubernetes Service and IBM Cloudant bring.
• Over time, the original apps that were initially moved to the cloud were restructured into cooperative microservices that run on IBM Cloud Kubernetes Service. This action further sped up development and time to deploy and allowed more innovation due to the relative ease of experimentation. They also released innovative apps with newer versions of microservices to take advantage of market and business conditions (that is, so called situational apps and microservices).

Context

Streamlining developer productivity and deploying AI tools to partners 4 times faster**

• Disruption is rampant in the payment industry that also has dramatic growth both in the consumer and business-to-business segments. But updates to payment tools were slow.
• Cognitive features are needed to target fraudulent transactions in new and faster ways.
• With the growing numbers of partners and their transactions, the tool traffic increases, but their infrastructure budget needs to decrease, by maximizing efficiency of the resources.
• Their technical debt is growing, not shrinking from an inability to release quality software to keep up with market demands.
• Capital expense budgets are under tight control, and IT feels they don't have the budget or staff to create the testing and staging landscapes with their in-house systems.
• Security is increasingly a primary concern, and this concern only adds to the delivery burden that all cause even more delays.

Solution

The Development Exec is faced with many challenges in the dynamic payment industry. Regulations, consumer behaviors, fraud, competitors, and market infrastructures are all rapidly evolving. Fast-paced development is essential to being part of the future payment world.

Their business model is to provide payment tools to business partners, so they can help these financial institutions and other organizations deliver security-rich digital payment experiences.

They need a solution that helps the Developers and their business partners:

• FRONT-END TO PAYMENT TOOLS: fee systems, payment tracking including cross-border, regulatory compliance, biometrics, remittance, and more
• REGULATION-SPECIFIC FEATURES: each country has unique regulations so that the overall toolset might look similar but show country-specific benefits
• DEVELOPER-FRIENDLY TOOLS that accelerate roll-out of features and bug fixes
• FRAUD DETECTION AS A SERVICE (FDaaS) uses Watson AI to keep one step ahead of frequent and growing fraudulent actions

Compute, DevOps tools, and AI that run in public cloud with access to back-end payment systems. Implement a CI/CD process to dramatically shorten delivery cycles.

Technical solution:

• IBM Cloud Kubernetes Service
• IBM Cloud® Continuous Delivery
• IBM Cloud Logging and Monitoring
• IBM Watson for Financial Services
• IBM Cloud App ID

They started by containerizing the payment tool VMs and putting them in the cloud. In a flash, their hardware headaches went away. They were able to easily design Kubernetes clusters to fit their CPU, RAM, storage, and security needs. And when their payment tools need to change, they can add or shrink compute without expensive and slow hardware purchases.

With IBM Cloud Kubernetes Service horizontal scaling, their apps scale with the growing number of partners, resulting in faster growth. IBM Cloud Kubernetes Service provides elastic compute resources around the world that are secure for full usage of modern compute resources.

Accelerated development is a key win for the Exec. With the use of modern containers, Developers can experiment easily in the languages of their choice, pushing changes to Development and Test systems, scaled out on separate clusters. Those pushes were automated with open toolchains and IBM Cloud® Continuous Delivery. No longer were updates to the tools languishing in slow, error-prone build processes. They can deliver incremental updates to their tools, daily or even more frequently.

Also, logging and monitoring for the tools, especially where they used Watson AI, rapidly integrate into the system. Developers don’t waste time building complex logging systems, just to be able to troubleshoot their live systems. A key factor for less staffing costs is that IBM manages Kubernetes, so the Developers can focus on better payment tools.

Security first: With bare metal for IBM Cloud Kubernetes Service, the sensitive payment tools now have familiar isolation but within the flexibility of public cloud. Scans for vulnerabilities are run continuously.

Results

• Lifting the existing monolithic VMs into cloud-hosted containers was a first step that allowed the Development Exec to save on capital and operations costs.
• With infrastructure management taken care of by IBM, the Development team was freed up to deliver updates 10 times daily.
• In parallel, the provider implemented simple time-boxed iterations to get a handle on the existing technical debt.
• With the number of transactions they process, they can scale their operations exponentially.
• At the same time, new fraud analysis with Watson increased the speed of detection and prevention, reducing fraud 4 times more than the region’s average.