Cloud
Leadership Series
Article #3

Many large financial institutions have a legacy of technical debt resulting from a landscape of complex interconnected banking and trading systems, constrained on-premises hardware, and a recent history of needing to prioritise regulatory changes. In the Capital Markets domain, vendor software upgrades have been deprioritised, or else deferred.

 

Current on-premises environments and legacy estates are prohibitive and prevent organisations from responding quickly to market conditions. The implications of this are far reaching, affecting everything from security vulnerabilities to being able to keep pace with the competition.

 

In contrast, cloud computing enables companies to instil that most critical of characteristics: agility. In providing a broader range of compute services over the internet, cloud providers can assist with the reduction of operational overheads. Importantly, this allows organisations to focus time and money on areas that generate commercial benefit and align with strategic business goals. The modernisation of an application for cloud adoption often brings a fundamental shift in how the system is designed, and organisations can start to realise the benefits that come baked-in with increased agility and scalability.

 

When it comes to DEV and TEST, there is no longer a lengthy lead time for environment provision, nor is there a struggle for capacity on existing environments. With infrastructure-as-code, such as Amazon Web Services (AWS) CloudFormation developers can safely spin up repeatable and reliable infrastructure for testing and spiral it down once complete – all while adhering to enterprise policies. This helps Capital Markets organisations become more flexible in their development approach.

 

Where production environments are concerned, the ability to scale dynamically is crucial. Essentially, it eliminates the need for operations staff to manually respond to periods of high market activity, batch re-runs or server outages. Scaling can also help with cost optimisation as it automatically matches supply with demand, reducing the ‘over-provisioned and under-utilised’ conundrum.

 

For Capital Markets firms, reliability and disaster recovery are critical. With internal and external SLAs and KPIs to comply with, cloud architectures need to be designed with reliability and fault tolerance in mind. With AWS, a large portion of these services are built with these foundational requirements already incorporated. Perhaps more pertinently, they can be tweaked and configured to meet the exact needs of the customer, replacing traditional failover architectures with something more fluid and adaptable.

 

Money always talks and, in banking and capital markets, cost is king. Application workloads must be constantly analysed to identify the most cost-effective method for each particular use case. The AWS Cost Optimisation whitepaper provides guidance on implementation, with a control on costs, delivering much-needed financial oversight and sustainability.

 

In Focus – Fundamental Review of the Trading Book

 

 

Regulatory changes are making cloud adoption more relevant and appealing than ever before for Capital Markets firms.

 

The Fundamental Review of the Trading Book (FRTB) is one such change. FRTB is accelerating the pace of cloud adoption, capturing tail risks by reporting expected shortfall defining the banking and trading book. This new definition confirms which assets (frequently traded vs. held-to-maturity) make it into capital calculations, eliminating ‘regulatory arbitrage’ between trading and banking books. Calculations take place in two ways: Simple or Internal Method Model (IMM).

 

Simple calculation is standardised, but the IMM approach requires banks to provide sufficient back testing evidence for the regulator. In terms of technology components, this delivers a requirement for extra compute capacity.

 

High-Performance Computing (HPC) grids currently provide the compute capacity for managing risk within capital markets. HPC grids use schedulers and orchestration tools to route business calculations to compute infrastructure and return results – and will continue to be a key component in regulatory compliance going forward.

 

For FRTB there will be a requirement for more calculations, hence more compute infrastructure, though the running of daily workloads, batch and intra-day pricing makes it difficult to accurately estimate the level of extra capacity required for back testing.

 

Purchasing extra compute hardware is an option, but it does come with its own complications:

 

• HPC managers will need to estimate the extra capacity needed for back testing.
• If under-utilised, extra costs will be incurred.
• Lead times for server deployment can be months, during which time new information may skew
  the initial capacity request.

 

Cloud, meanwhile, offers the ideal alternative to finding that extra capacity: the ability to scale up compute capacity as required, and scale down when not. Gone are lengthy lead times and estimation headaches. Now, servers can be deployed within minutes. By modernising and extending the HPC grid footprint to the cloud, organisations can begin working towards FRTB compliance. And just as importantly, they can be decommissioned immediately after testing to manage or reduce costs.

 

In addition to AWS’s ability to provide extra compute capacity, there are subtle elements that can be tweaked to ensure you get the best value when running computations on AWS:

 

• For stateless workloads, spot instances are ideal resources. These are compute instances at discounted prices and when running 1000s of calculations, the savings can become significant.
• HPC environments need compute capacity. However, they do not necessarily need the latest CPU. By selecting a previous generation instance type, you can achieve lower costs and there is typically more availability, particularly with spot instances.
• FSx for Lustre (Amazon FSx for Lustre) is a specialised storage solution. It is scalable and provides high performance throughput for compute instances.
• AWS also provides solutions to meet network throughput requirements such as Enhanced Networking and Elastic Fabric Adapter (which supports OS bypass).

 

Traditionally, HPC environments have been quick wins for capital markets to migrate to the cloud. In fact, for many they have been, or will be, one of the first projects to be migrated, and AWS has worked with some of the main orchestration tools and HPC vendors to validate and test their products on the cloud. As a result, many organisations are now utilising AWS for their HPC environment, either as burst capacity or for larger portions of their estate.

 

Whilst HPC may be the first step onto the cloud, it shouldn’t and doesn’t stop there. The cloud provides ample services to compete with most on-premises solutions, and if it doesn’t now – it will.

 

Journey to cloud adaption

 

The pace and extent of cloud adoption is being driven by a number of factors, from regulatory changes to the size and scope of individual firms. By modernising the IT landscape through cloud adoption, organisations can start to realise true business benefit and react to new regulatory requirements.

 

Size in particular plays a key role in the pace of migration; the investment banks and asset managers we mentioned previously migrate at a more cautious pace than smaller firms, who are often opting for cloud-hosted SaaS or PaaS models. With less operational overhead, this allows them to focus on achieving a competitive advantage through the smart application of new technology.

 

An organisation’s agile maturity is also salient. DevOps and cloud are key to unlocking the true value of agile delivery practices. Those organisations that are more advanced on their agile journey are keen to embrace the full cloud and DevOps benefits.

 

Finally, trigger events, such as hardware or software EOL horizons, force organisations to invest budget into the cloud. Organisations are seizing the opportunity to invest in cloud rather than pay for bespoke support contracts.

 

Cloud control

 

 

The modernisation of environments and technology stacks means that Capital Markets firms can respond more quickly than ever to changing business and regulatory needs. Cloud provides many of the services needed to help organisations along this journey. At its very core it focuses on classic IT services to replicate on-premises environments: compute, storage, network, database. Once organisations make the commitment to adopt cloud and modernise their estate, the options begin to grow and the range of services, in particular managed services, starts to become a ‘tangible’ reality. While the road to cloud is well underway for Capital Markets, the pace and velocity is likely to increase due to the increasing regulatory demands as well as the constant search for efficiencies and profitability.

 

AWS has been at the forefront of helping transform the financial services industry by enabling firms to modernise disparate legacy systems for improved agility and scale, meet rapidly changing consumer behaviours and expectations, drive business growth by harnessing data and innovation, and build with confidence on the most secure and resilient cloud.

 

AWS does this by providing CMs the secure, resilient global infrastructure, services, and expertise they need to accelerate release cycles, improve decision making, reduce costs, and streamline operations. The infrastructure delivers the highest level of security and resilience to meet local and global compliance regulations while supporting innovation at scale and speed.

 

Through readily available services in high-performance computing, data and analytics, and machine learning, AWS makes it easy and cost effective for financial services to realise the benefits of digital transformation. These ideas could involve new products and services, or they could involve creative new risk management solutions.

 

DXC Luxoft works with over 80 clients globally and is a market leader in complex capital markets software packages. We spend time with clients to understand their architectures, requirements and concerns and, together with AWS, offer bespoke services to help clients of all sizes modernise and migrate their services to the cloud with minimal disruption.

 

DXC Luxoft has vast experience within HPC and have worked with many of its clients to migrate risk and rates HPC platforms to the cloud, and continue to provide a managed service for many of these clients. Working alongside architecture managers and business stakeholders, DXC Luxoft has delivered proof of concepts through to production deployments allowing for scalability and a reduction in compute time, all whilst achieving an overall reduction in infrastructure costs. In a recent project, DXC Luxoft and customer achieved a 50% reduction in compute time when benchmarked against existing infrastructure. Throughout these engagements DXC Luxoft has worked closely with the cloud governance and CISO teams. This ensures its solution meets enterprise standards and has led to widescale adoption with other HPC applications following a simple, pre-defined framework.