By Anthony Amicangioli, Founder and CEO, HPR
The universe of high-performance capital markets infrastructure is highly fragmented today, with many trading stacks built on stitched-together legacy software. Typically assembled from a combination of in-house and outsourced solutions with a mash-up of programming languages and design philosophies, these stacks can struggle with spotty performance, difficult upgrades, and uncertain interoperability.
Based on decades of experience as a founder, capital markets CTO, and designer of high-performance networking systems, my conclusion is that these challenges can be solved by providers willing to adopt a comprehensive, true platform—built through a hyper-unified approach to development.
For market participants looking to address critical needs like pre-trade risk, enterprise trading risk, market access, high-speed market data delivery and consumption, or matching engine technology, a platform-as-a-service solution seamlessly integrates and unifies all solutions across these firms and product sets.
This is the approach we have taken since I launched HPR in 2011. Through rigorous engineering discipline, it can deliver unmatched reliability, scalability, and performance in all areas of trading infrastructure.
Advantages of a True Platform
The use of unified platforms in the development and deployment of complex solutions has proved successful across a number of major industries, including technology, healthcare and telecommunications. It offers advantages that ultimately result in greater power and efficiency for users.
Platforms can dramatically speed up the development of new products and features. Working from a singular codebase allows engineers to immediately leverage new capabilities across an entire product suite. Code can be reused extensively, reducing bugs and incompatibilities while promoting greater agility and collaboration across teams. This also makes maintenance and upgrades far simpler, as updates to shared code are deployed in one stroke across an entire ecosystem.
Other benefits include a unified and improved client experience, as there is no need to learn different user interfaces or workflows for each application. Data is aggregated across the platform, enabling seamless analysis and insights. Greater scalability and reach are also facilitated, as are systemwide security and compliance.
Google and Amazon, two of the biggest names in technology, are widely regarded as pioneers in the use of platforms. Google’s diverse product suite, which includes search, Gmail, YouTube, Google Maps and Google Cloud, is built mostly on shared infrastructure and unified services. Central to its platform, Google largely maintains a monolithic source-code repository, an increasingly popular approach among leading technology companies.
Early on, Amazon implemented the renowned Bezos API Mandate, which required every team to connect through shared interfaces, breaking down silos, and driving integration at scale. This is often cited as central to Amazon’s ability to scale so dramatically, so quickly. AWS was born out of Amazon’s platform approach.
The Dangers of Fragmented Architectures
In contrast, the trading stacks operated by many capital markets firms today consist of products cobbled together from legacy software and written in multiple programming languages. Legacy software faces multiple downsides that can impact performance, including:
- Inconsistent and error-prone system designs often caused by turnover in the CTO seat and M&A activity
- Slower lead time in new feature implementation
- Performance degradation and maintenance challenges causing undue pressure on engineering resources
- Disparate applications leading to poor interoperability
- Higher integration or upgrade costs driven by stitched-together applications
Any of these issues can interfere with trading activity and be responsible for trading interruptions.
Lessons from a Comprehensive Approach
I have learned a great deal from building, creating, and operating a comprehensive, true platform, and am convinced today more than ever that this is the future of high-performance capital markets infrastructure.
The advantages of a platform approach stem from building on a singular, highly unified codebase that functions as the common language and fabric of the ecosystem. From such a codebase, it is possible to build a comprehensive, centrally controlled suite of applications that work in concert across the high-performance trading universe.
For example, a broker-dealer may need to access a broad range of trading venues as well as enable ultra-low-latency direct market access and gain access to low-latency, deterministic market data. A true platform can easily meet those varied needs. Because these solutions are designed to operate together, many of the friction points commonly experienced in legacy software would be eliminated.
Such an approach allows low latency, throughput, and determinism to be prioritized in a world where nanoseconds matter. Inefficiencies caused by disparate applications don’t exist. It becomes far easier for clients to scale on a true platform, accommodating greater trading volume as well as adding new services and markets. Importantly, new features can be rolled out quickly and uniformly.
In the world of high-performance trading, a true platform stands poised to drive the future of capital markets.
