By Himanshu Gupta, Senior Architect, Capital Markets, Solace
The data movement of every trade unlocks a competitive edge
Prices in financial markets can change in the blink of an eye, so the ability to execute trades faster than competitors unlocks huge competitive edge. That demands a high functioning tech stack and an infrastructure to match a 24×6 global trading system. Himanshu Gupta, Senior Architect, Capital Markets at Solace, knows the issues only too well. Previously a q/kdb+ developer for a major bank and data engineer at a capital management firm, he knows first-hand that infrastructure is a critical differentiator in high performance hedge fund trading.
Very few industries need the sort of high-functioning technology stack that is required in capital markets. But the infrastructure that supports this tech-stack is crucial to success – in hedge fund trading in particular, the need for speed and integrity of data is paramount.
A hedge fund IT infrastructure must be able to distribute messages quickly and effectively to the necessary processes such as from the market data feed handlers to algo trading applications to order management system (OMS) and to finally an execution management system (EMS). As the messages hop from one process to another, leading to delay, every millisecond or microsecond that can be salvaged makes a difference to the bottom line.
However, even today, many hedge funds are struggling to modernize their legacy message solutions. While some are stuck with legacy inferior solutions that are too difficult to untangle, others have adopted different solutions over the last few years leading to non-standardized mess. As a result, several hedge funds use a variety of messaging solutions that have not been designed to meet the needs of the industry from a speed, efficiency, resilience, and security perspective.
There are four key features and considerations that a messaging solution that will shave off microseconds from a transaction in a Buy or a Sell transaction that will provide hedge funds with competitive advantage.
1. Performance – market data distribution with high volume and minimum latency.
Market data is high volume and extremely volatile – hedge funds receive up to microsecond data feeds from sources such as Refinitiv and Bloomberg. Once market data is consumed, it is fanned-out to downstream consumers such as Risk, PnL, OMS, and EMS systems.
The message volumes can vary significantly day-to-day but depending on the number of securities a hedge fund is trading and the global markets it’s participating in, these volumes can be in millions of messages per second. A performant message broker must be capable of handling such high volumes and be able to scale horizontally as the volumes continue to increase.
Additionally, these messages need to be delivered with minimum latency. For example, to achieve latency as low as 18 microseconds provided by Solace PubSub+ brokers, messages need to be delivered in-memory without being persisted to disk. This comes at the risk of potentially losing some messages but that’s acceptable for market data distribution use case which requires at-most-once quality of service.
That’s what gives an edge in capital markets environment – when you’re making your decisions, you need to have the latest view of the market, the exchanges, and what’s happening to the stock price. Consider how quickly FX markets can move due to macro events such as key inflation reports being released.
FX market moves at speed
I recently encountered a scenario within FX trading, where traders saw a spike in market data volume due to release of US Consumer Price Index – hugely important when providing the service of trading these currencies for customers. When you’re buying or selling a certain currency pair at a certain price, you need to have the latest view of the market in an instant.
If you’re dealing with stale information, depending on which security you’re dealing with, and how often it trades, even a millisecond or few microseconds delay can have huge consequences. Speed is of the essence in hedge fund trading, and a good tech stack is nothing without architecting efficient data distribution to minimize latency in market data use cases.
2. Trade order distribution – with maximum throughput and zero message loss
Unlike market data distribution, trade order distribution use case requires at-least-once quality of service. Trade order distribution spans the pre- and post-trade life cycle and ranges from origination of trade ideas to trade execution to post-trade settlement and reporting.
Message volumes of trade orders are much lower than market data but every message needs to be securely delivered. To achieve this quality of service, message brokers need to leverage persistence and acknowledgements.
As messages are delivered from publisher to broker to consumer, acknowledgements are shared to ensure message delivery is guaranteed which ensures a hedge fund’s trading stack will never lose a message. The additional overhead means the latency will be higher than at-most-once quality of service but that’s acceptable given the additional assurance of delivery. Not many brokers can support both quality-of-service which leads to some hedge funds using different brokers and over complicating their technology stack.
As an example, one of the world’s largest hedge funds uses an event-enabled solution to distribute trader order updates published by local instances of their OMS over a global event mesh. The order updates from regions such as New York, London, Hong Kong, Singapore and Tokyo are consolidated back to its primary data center for risk analysis of their positions in real-time.
3. Trading markets are global: ‘Always on’ operations are critical – with high resilience and robustness
While speed and performance are of utmost importance in Capital Markets, a critical requirement for hedge funds is to have a message broker that is robust, reliable, and resilient.
Everything in capital markets requires high availability and, when required, disaster recovery. Many hedge funds operate 24×6 – especially those participating in global equities and commodities markets. If one broker goes down because of any issue, the trading system will come to a halt and lead to potential capital losses. Even an outage of 30 seconds can be extremely damaging to a hedge fund’s operations and reputation depending on the trading strategies it utilizes.
Round the clock availability, no matter the weather!
Resiliency is key here particularly when it comes to dealing with unforeseen events, such as Hurricane Sandy in New York couple of years ago Hedge funds need to continue trading. From a regulatory perspective, it is also important that critical trading systems have disaster recovery strategies. Additionally, many large investors such as pension funds, prefer to invest with hedge funds that have proper high availability and disaster recovery procedures.
Hedge funds need to be able to tolerate not only loss of a single broker in a data center, but also an entire data center becoming unavailable. To achieve such a high level of resiliency, hedge funds need to deploy two brokers in the same data center for high availability, and then two more brokers in a different data center for disaster recovery. These strategies are key to ensuring business continuity.
4. Meet the virtual broker – keeping data exchange secure – internally and externally
It’s not just sensitive financial information that needs to be secure. Within hedge funds it’s the internal data on client orders that can become very sensitive. Many hedge funds use portfolio managers (PMs) – who bring their own teams, expertise and specializations.
But this gives rise to internal competition, so data needs to be communicated separately. When important trade order messages are being distributed over message brokers they need to be encrypted and secured.
One virtual broker at a time
This is where virtualization comes in to keep messages separate. Hedge funds can dedicate one virtual broker to a PM, another virtual broker to another PM and the messages will be separate. This applies to different teams within a hedge fund as well, such as PnL and Risk. Different teams should be able to use the same underlying broker without interfering with each other’s operations. One team should not be able to access another team’s messages unless authorized. This also helps drive the overall cost down for the middleware team by using the same underlying broker to serve multiple teams.
Someone who has access to one virtual broker cannot connect into the other virtual broker and access that data, all connections and data in motion can be encrypted as well. Additionally, to further enforce strict authorization, special Access Control Lists (ACLs) can be utilized to enforce which users can publish or subscriber to which data.
IT architecture at the core of hedge fund success
There is a fitting quote that summarizes hedge fund trading from Hendrith Vanlon Smith Jr: “to hedge effectively, you have to reject the silo mentality and instead embrace the systems thinking mentality.” This applies even down to the IT architecture level for hedge fund organizations and quicker, complete, robust, and more secure data movement can differentiate every trade – in an industry where every microsecond counts.
