Today, you need to know what’s happening in your business in real-time. Faced with pressure from legislation and the emergence of new digital ecosystems and partnerships, traditional banking incumbents compete with new fintech players across the entire financial services arena. The value chains that previously existed are starting to unbundle based on how customer data can be used. Alongside this, the COVID-19 pandemic continues to create more economic uncertainty, affecting how organizations reset and renew their digital strategies. 

Harnessing all your data in real-time is a great objective, but achieving this at scale and with the right economic model is a harder challenge. But leveraging real-time data can support a range of business goals - from higher levels of reliability and resilience for your applications through to better customer experience. Understanding your digital journey is the first step around making data more valuable, and it requires you to manage the intelligence gaps that exist.

Dealing with this means changing your approach. According to Ben Hunter, Financial Services Sales Director, UK, Sumo Logic, “There is a huge opportunity for fintech companies to take more market share away from the larger, more complex organisations that are struggling to serve a changing, growing digital customer base. In some cases, larger organisations have already realised they need to innovate themselves. They made the decision to create subsidiary companies or to partner with fintech companies directly that are better equipped to act on their behalf. Whatever market you are in, you have to think and act differently.”

Containers and cloud

Thinking and acting differently means looking at new ways to run your systems. From an infrastructure perspective, there are two big changes in how companies develop and operate their applications. The first of these is software containers, which can host application components as small, self-contained units. Containers can be used to support specific services, and more containers can be added whenever you need to scale up. A container orchestration platform like Kubernetes can automate this process in response to demand from customers.

Banks tend to be at the forefront of new technology deployments, particularly where these implementations can support faster time to market and speed to value. The growth adoption of Kubernetes correlates strongly with multi-cloud adoption too, particularly for companies deploying agnostic interoperable layers for Service Mesh, PaaS and IaaS mixture of cloud workloads. Multi-cloud is something that is currently beyond most banks. In contrast, fintech companies are keen on multi-cloud from the start.

What can you take from this?  If you aren’t already making use of containers, then be prepared to adopt this approach and technology in the near future. Using modern APIs, applications and microservices componentry can also help your team deliver faster and reliable user experiences, so be prepared in advance.

Serverless and building faster

The second big trend around infrastructure is serverless computing. Alongside containers, serverless computing and functions as a service are being adopted more and more to support modern application architectures. For fintech companies, this approach is popular as it helps you get services up and running even faster, there is no engineering overhead for managing that infrastructure, and you only pay for what you consume. 

The move to serverless is interesting as developers think this can remove a huge swathe of infrastructure management overheads for them. However, for developers - and for banking and financial services teams in particular - there is still a definite demand for data around these applications. Monitoring and troubleshooting your serverless implementation is essential for four reasons:

• To improve and optimize your software development lifecycle, or SDLC - serverless abstracts a lot of complexity out for you, but that complexity does not disappear.  You have to track how well your app is performing over time, so that you can carry on hitting availability and performance targets and take care of any SDLC issues.

• To track and optimise your budget in real-time - as serverless is a ‘pay per use’ model, spikes in demand for your application can become costly very quickly. Having a good economic model for your service is essential, so that any additional costs for serverless over your estimates directly lead to revenue. If you can’t do this then managing calls so costs don’t escalate quickly is just as important.

• To spot latency and performance issues - any performance and latency issues due to code testing and delivery problems can affect your spend too as you scale. Observability data is therefore required to control your costs and make the most of any spend/investment.

• To increase your security and compliance – alert fatigue, new issues get discovered all the time, so any threat, vulnerability or anomaly that occurs in your serverless application has to be found quickly and resolved. Centralized Observability telemetry provides you with the trusted, right context and insights to predict, detect, investigate, resolve and recover from the issues in highly regulated environments. 

The adoption and acceleration of serverless helps you innovate and reduce your costs around application delivery. However, managing cultural change is harder to embrace with a traditional mindset in place. Getting observability data on how your application performs and scales over time when you use serverless and container architectures is essential. You should also put processes in place for how you make decisions around managing applications when your application components are highly ephemeral and immutable. 

Getting the right data at the right time

Getting data is essential, whether you are in a bank or in a fintech. Without good data sources, you won’t be able to build the services that customers value. 

Running this requires data that you can understand and use as fast as any transactions or interactions with customers are created. By centralising all logs, events, traces and metrics data that your applications and infrastructure generate, you can build that set of data and deliver a single source of truth for the business as a whole. This can then be turned into actionable intelligence by supporting more AI and Machine Learning use cases over time. Using tools like dashboards, alerts, and real-time streaming and pattern detection, we reduce diagnosis and troubleshooting time and make better decisions with context, on what to do next.