Moving to the cloud – it’s a risky business

Thursday, May 13, 2021 | By Leo Armer, Senior Director, Pre-Sales, SS&C Algorithmics, and Rob Seidman, Director Cloud Innovations, SS&C Algorithmics

Moving to the cloud – it’s a risky business

Technology and IT management have become second nature for many financial firms. There is no shortage of anecdotes from the last decade where C-Suite executives anticipated the changes. One such CEO of a Tier 1 European Bank back in 2015 foretold, "In the future, we will be a software company.” Our view as a software company, was we felt confident there would continue to be a greater reliance on IT infrastructure. The emphasis on infrastructure remains, although the business challenges are evolving.

Many financial firms have spent years—cumulative centuries of person hours—designing and developing their own software components and approaches to risk management. They may have invested with a vendor or developed their own technology stack in-house, but in almost every case there is customization. From developing complex pricing models, insurance modeling and credit processes to collateral management, treasury cashflow systems, finance, algorithmic trading and trading book risk sensitivities, projects were undertaken and justified because of the “edge” they gave the institution over the competition.

We’re now in 2021, where a global pandemic and political instability have brought about significant volatility in the markets; this is coupled with years of financial regulation and a management and regulatory paradigm where the cost of running these (or any) financial applications is closely scrutinized. When looking at the latest risk news, it often seems that market volatility is pushing organizations to run many more possible risk scenarios. Many regulations are asking for greater detail and insight, and in many cases, an advantage can be gained through increased granularity or improving transparency in the way risks are managed. This blog does not intend to cover all possible drivers and motivations in moving to the cloud, we see value in sharing some key themes that have been observed. During a recent webinar on this subject, participants discussed:

  • Business demands that we increase the number of calculations/simulations.
  • Users require the ability to run on-demand, what-if requests.
  • Total cost of ownership is a key consideration, looking for cheaper and resource-efficient solutions.

Cloud technology ticks all of these boxes—it’s a way to make the software cheaper to run, supports the ability to scale, and provides flexibility with payloads. As the decision to move to the cloud often feels like a big leap of faith, what options do you have to get there smoothly, without disruption to business as usual?

You might be put in the envious position of working out the difference between one cloud supplier and another. Or, maybe your organization has made life easy and set a corporate standard to use a specific vendor. Either way, the key cost drivers of financial risk management remain the same—risk solutions, unlike some other financial applications, can be very compute-intensive. This is further amplified if you need to run intense scenario simulations on a batch basis. In addition to this, some applications are also very data-intensive, meaning they need a significant amount of scenario data and pricing/modeling data in order to run. If we look at the amount of time traditional on-premise hardware takes to perform risk calculation, the results can be quite staggering. For more than half of our clients, the CPU utilization time of the servers used for risk management workflows is less than 25% of overall uptime. This means that the machines are sitting around doing nothing 75% of the time!

Utilizing the elastic nature of cloud-based servers yields a significant reduction in this wastage and makes the applications more efficient. Risk solutions need to also be agile; when a stress test scenario is devised or a major market/counterparty credit volatility event takes place, the business may want to run multiple new scenarios. How long would this normally take? Is it hours/days? Or is it minutes? An elastic cloud environment should enable access to servers on-demand and provision hardware when you need it. To your business, the quicker you can run this the quicker you can respond. The cloud paradigm of elasticity has at once introduced both cost savings and increased business agility.

The cloud can also help with developing new models, innovation and new business processes. We recently ran a cloud workshop with one of our clients where the client wanted to understand how to upgrade from one version of a pricing model to another. The challenge was that they needed to create a separate User-Acceptance Testing (UAT) environment for testing with 100+ days of market data to run with one model and then repeat the same process for the other. It would take about one day to run and two more days in total to fully analyze. When we used the models loaded in separate containers—a cloud software technology designed to improve portability and standardization—running on the cloud the calculation and analysis took less than two hours. The excitement coming from our client was palpable; they realized this technology had the power to change their business model.

In order to utilize the cloud, you will need to get data to it. In risk, as we have mentioned, this can be a substantial amount of data. Risk management relies on both internal data as well as vendor-supplied data. The quality and availability of this data is central to many front and middle-office functions. Data egress and ingress is one of the key cost, regulatory, and operational concerns around cloud adoption. Each firm or department needs to assess how much data needs to be uploaded (ingressed to the cloud platform) or downloaded (egressed from the cloud platform), and does that outweigh the financial benefits of moving to the cloud? There are cases where data transfer is both punitive in cost and time, but as more applications start to move to the same cloud infrastructure, the less pressing this issue becomes. This business challenge is often referred to as “data gravity”—as data has cost to move around, data gravity is the effect of data to attract both other data and data-consuming applications to where the data itself resides. In time, as we see more applications on the cloud, the problem effectively disappears, perhaps even incurring a larger cost from remaining on-premise.

The decision to move to the cloud ultimately will come down to whether your risk applications are ready to take advantage of the cloud infrastructure. Existing systems may have to be re-architected before an inflection point of profitability is achieved by moving to the cloud. The promise of cost reduction and the potential for increased agility of business processes are the key areas that will affect the decision-making on the future of financial risk management technology infrastructure. Moving to the cloud can certainly be phased and will take some time for many financial institutions, but with the right structure and right approach to the discussion, the benefits become clear.

Visit our Algorithmics Cloud Platform product page to learn more.

Alternative Investments, Asset Management, Insurance, Risk Management, Wealth Management

Theme picker