Firms Struggle with Cloud Migration

As the move to the cloud gathers pace, capital markets firms are facing growing pressure to transition their front-office operations quickly, effectively and cost-efficiently to the cloud, according to Matt Barrett, co-founder and CEO at Adaptive. 

Matt Barrett

“The challenge is that many on-premises legacy systems aren’t easily transferable to cloud infrastructure and require either a substantial re-engineering effort or a complete rebuild,” he told Traders Magazine.

“Whilst firms are once again faced with the age-old dilemma of buying ‘off-the-peg’ products vs building bespoke tech solutions, the cloud is reshaping the debate. Whilst the answer is now clearly ‘hybrid’, the question has become: at what layer of the tech stack do you switch from ‘buy’ to ‘build’,” he said.

According to Barrett, the cloud enables and requires firms to think differently about buy vs build, and the layer of abstraction they choose to build upon.

When asked if a hybrid model is possible, Barrett said that the word ‘hybrid’ starts to become overloaded when discussing a migration of liquidity to the cloud. 

The more common usage of the word ‘hybrid’ was in response to the buy vs build choice, intended to represent a more sophisticated set of decisions around when and what to build or buy, he said.

“As venues and liquidity move to the cloud, market participants that connect to a range of venues will find themselves, for some time, in another hybrid world,” he said.

“This world, of some venues existing in the cloud, and some remaining in physical co-lo facilities, requires another sophisticated response,” he added.

 To navigate this hybrid landscape, and the piecemeal migration from a primarily co-lo world to a primarily cloud world, firms must choose the right underlying technology to build their front office technology, Barrett said. 

This technology must be equally capable to be deployed on-prem or in a co-lo as it is in the cloud, he said, adding that it must provide a sufficiently strong abstraction layer over the underlying network infrastructure that, for performance and resilience reasons, any system built on top does not need to be re-engineered as it migrates from physical to cloud infrastructure.

“Our belief is that Aeron, the open-source messaging and clustering technology, provides exactly this abstraction layer,” he said.

The pros and cons of buying vs building 

Buying ‘off-the-peg’ vendor products facilitates a quick and cost-efficient deployment of operations into the cloud – allowing firms to not lose ground to competitors during the shift to the cloud, at least not in the short-term, commented  Barrett.

“But truly cloud-native solutions – where deployment, integration, and operation are all streamlined for the cloud – are usually deployed as vendor-hosted SaaS solutions,” he said.

“Front office latency requirements necessitate a thoughtful design to ensure a suitable end-to-end performance. Integrating SaaS APIs, hosted in other cloud regions, can add latency and jitter,” he added.

Barrett further said that building bespoke cloud-native tech solutions allows firms to build differentiated trading technology from the ground-up, tailored to their specific needs, whilst retaining control of the management and pace of innovation of their tech stack. 

Nimbler, they can respond more quickly to the changing market conditions by redesigning and customizing their technology as they see fit, he said, adding they can also ensure their system design minimizes end-to-end latency and maximizes resilience, critical in the front office.

“There is however a cost factor,” Barrett stressed. 

“In the short-term, building tech stacks from scratch requires a much larger investment, both in time and resources. Over the long-term, the investment pays off by empowering firms to bypass the cookie-cutter tech trap,” he said.

Barrett noted that the cloud was not built for capital markets, and it is unlikely to be (substantively) re-engineered for it. 

It was initially created for retail e-commerce, he explained, saying this means that there is currently a significant capability gap between what capital markets infrastructure needs – like multicast, storage, reliability – and what the cloud can provide. 

However, this gap is continually and quickly getting smaller, according to Barrett.

As cloud platforms’ technical limitations are being addressed, firms have been leveraging cloud-native messaging solutions that enable low-latency, high-throughput and fault-tolerant trading technology to build and deploy highly performing and resilient capital markets infrastructure, he said.

Barrett said that the move to the cloud will significantly disrupt the telecommunications, co-location and hardware ecosystem through which market participants connect to and trade with each other. 

“With venues moving to the cloud, the marginal cost of venue connectivity will tend to zero, as the cloud makes interconnectivity simple and free. Firms handling market data will find themselves competing with cloud providers at rock-bottom prices for data management and storage,” he said.

“No longer burdened with high trade execution and data storage costs associated with current physical infrastructure, there will be an accelerating shift in capital markets firms’ mindset and approach to their tech stacks, with considerable attention redirected towards innovation,” he added.

According to Barrett, resources will be reallocated towards building and owning bespoke trading infrastructures, before outsourcing the day-to-day management of these infrastructures to a trusted third-party. In this scenario, firms benefit from the best of both worlds – striking a balance between preserving their autonomy in driving innovation and capitalizing on external expertise to ensure the seamless execution of operations post-build.

“This mindset change has the potential to trigger a seismic shift that will ripple through the entire financial ecosystem,” he stressed. 

“Fuelled by the emergence of innovative differentiators and increased competition, access to financial products will be democratized and new AI- and ML-based trading algorithms and strategies will surface – irreversibly transforming the nature of capital markets,” he concluded.