By ArunKumar TR – Director, Technology, Synechron
Challenges due to data growth:
With data growing at a rapid pace, companies need to scale their infrastructure to handle and process huge volumes of data. Especially for financial institutions like Banking, Insurance, Capital Markets, Trading platforms generate large volumes of transactional data and companies need to persist historical data for efficient processing to meet the ever-changing functionalities.
According to a study of IBM in 2015, it is estimated that every day we create 2.5 quintillion bytes of data and that 90% of the data in the world today has been created in the last 2 years. As many financial firms and other industries going digital, The amount of Non-cash transactions has increased by 14% to reach 708.8 Billion transactions across globe in 2018-19 as per World Payment report 2020.
The sales of smartphone and handheld devices has quadrupled in 2018-2019 the number of transactions on e-commerce channels has significantly gone up at the same time leveraging the innovation in the digital payment industry. At NASDAQ 24 Million trades are executed per day generates enormous volumes of data.
Handling the challenge:
To handle such Real-time as well as offline applications to be stable and processing such huge volumes of data, it requires quite a bit of effort to setup and maintain an infrastructure on-premise. All these scenarios have led financial institutions to re-strategize their infrastructure capabilities and offload or on-board some of their processing to a Hybrid cloud strategy.
As per Gartner, By 2023, the leading cloud service providers will have a distributed ATM-like presence to serve a subset of their services. As the analytical demand is growing to adapt to millennials requirements, Financial institutions need to step up their infrastructure for advance analytical needs to cut down the processing times.
For Example, In the current day Commercial or Retail loan processing would take anywhere from couple of days to weeks depending on the type of loan. In order to cut down on the processing times and providing instantaneous approvals, they need to leverage latest cutting-edge AI/ML technologies to process numerous unstructured documents to ascertain a decision on the loan. With 100’s and 1000’s of such applications requests being logged every single day, it’s quite a challenge for the institutions to manage and process such requests. T
Apart from this example, there are hundreds of case studies that requires institutions to scale up the infrastructure on-premise. And just like any other scaling activity, scaling up infrastructure On-premise/Private Cloud is complicated with its own pros and cons.
Few considerations for Hybrid Cloud Scenarios:
Dynamic workloads: With new dimensions to business dynamics, a scalable infrastructure is required for the dynamic workloads on the applications. Public clouds can linearly scale to meet the dynamic workloads of the applications. But ensuring that sensitive data remains on private cloud or on-premise and secure within the network.
Big data processing: With complex analytics implementations increasing along with volume and variety of data, the big data infrastructure would come to rescue in terms of reducing storage cost and processing petabyte scale data. For use cases such as information discovery on Business performance, Clients insights and many other AI/ML scenarios, require a scalable infrastructure as the current on premise infrastructure would short fall the needs.
Incremental Cloud-Adoption: It’s not necessary that we need to go with the big bang approach to migrate all the required applications to Public cloud. Identifying applications where there are less/no sensitive data being processed on an application can be targeted for Public cloud movement and can be chosen to run in the hybrid cloud – with part of the application on-premise and part of it on public cloud. This will help the companies increase their cloud-adoption in a phased manner.
Another scenario would be modernization and scaling of legacy monolithic applications built over a period of time. Due to limitations in the architecture, the infrastructure technology of that time, or even the application design itself, this is another complex task to achieve. So, such applications can be considered as potential candidates for hybrid cloud use cases. With the rapid evolution of various technologies including microservices architecture, it is easier to identify such applications that may require considerable amount of refactoring and deploying them on Hybrid cloud gives organization the power to scale and manage their applications. The process of Identifying – Refactoring & deploying them can be iterative for several application components covering all the components of a monolith.
Temporary processing capacity: In the perspective of staying relevant by targeting Quick Time To Market (TTM), Organizations can leverage public cloud with managed service. This it reduces lot of effort compared to setting up infrastructure on-premise, as they offer wide range of servers from normal 8 core processing to high end GPU machines. Example: We can leverage 64 core GPU machine to train machine learning models and stop utilizing the GPU machine from your public cloud infrastructure once your models are trained.
Future Flexibility: Organizations which look for constant change and want to bring in innovative financial products to consumers, can benefit from Hybrid cloud implementations, by leveraging full potential of Hybrid cloud by deploying High availability software’s for internal use, especially the trading applications which is accessed by Broking firms, Traders , Investment bankers using various devices and connecting seamlessly and at the same time provide resiliency for their data. As petabyte scale data can be stored on-cloud and computed on-cloud, it gives a greater flexibility for applications to co-exist and be in sync on both on-premise and on public cloud.
As we have seen so far, leveraging the best of both worlds is the key to highly performant, scalable, and manageable applications on cloud. There are very less solutions that still gets designed to function solely on a public cloud solution or a private cloud solution. Why limiting our options? By adopting a hybrid cloud approach, we can tap the best of both worlds simultaneously, and yes, win the game !