As trading firms look to improve business performance, perhaps using sophisticated strategies running at speed, and regulators demand increased transparency and integrity across trading environments, the need is for tools that can monitor and analyse complex business flows. IntelligentTradingTechnology.com spoke to Velocimetrics’ CEO Steve Colwill about the need for real-time business insight and how it can be achieved.
Q: To start with, please can you introduce Velocimetrics and explain what it provides?
A: Velocimetrics provides real-time business insight to the global financial community through comprehensive monitoring and performance analysis.
Q: Can you detail the key functionality of Velocimetrics?
A: Velocimetrics enables firms to monitor their complete environment effectively, analyse their business activities and use this information to improve performance. Velocimetrics monitors traffic flows, tracking the paths of individual business items through applications and networks in real time. In addition to measuring and analysing hop-by-hop latency, Velocimetrics can alert on stalled or missing items as they traverse an environment, raising alerts as soon as degradation starts to occur.
By monitoring processes from a business, rather than technical infrastructure, perspective, users can analyse business-level information sliced and diced to provide the insight they are interested in. For instance, Velocimetrics enables performance to be examined at both the statistical and individual level for quotes, orders, instruments or trades, and a multitude of other metrics. This information can then be used to help the business identify how and where the firm is making money and, more importantly, where it is being lost.
Q: How have you developed the architecture of Velocimetrics to deliver this functionality?
A: To track complex business flows, Velocimetrics takes a multi-technology approach to instrumentation. In addition to the capture and analysis of network traffic, Velocimetrics provides adaptors that can capture information from log files, message buses and in-application code points. Once data is captured, it is streamed in real time to our scalable analysis cluster where data can be correlated and analysed. Correlation and cause-and-effect association between different classes of business items is supported by a simple to use point-and-click wizard, but we also allow customised correlation logic to be plugged into our analysis cluster via either a simple scripting language or Java code fragments. Rather than combining data capture and partial correlation/chaining in appliances, Velocimetrics separates the capture and analysis processes entirely, with the result that distributed environments can be monitored at much lower cost than would otherwise be the case.
Q: How does Velocimetrics differentiate from other business flow and transaction monitoring solutions?
A: One of Velocimetrics’ core strengths is its ability to deliver complete transparency across highly complex environments. Monitoring these environments using more traditional methods, such as monitoring at the institutional boundary, or on a system by system or network only basis, can often create blind spots. Velocimetrics takes a different approach, capturing data from numerous points and using multiple sources and techniques based on the client’s requirements. For example, this could include monitoring messages as they traverse middleware queues, conducting non-intrusive network data capture or embedding application agents. Velocimetrics is then able to provide clients with real-time visibility across their entire environment, supported by detailed business-level analytics.
Velocimetrics deals with the complexity of real-world systems by modelling multiple cause-and-effect relationships and complex correlation scenarios across numerous systems and monitoring points in real time. For example, a trade may be the result of a specific price quote that is, in turn, generated from multiple market data events. Such chains can be formulated through a mixture of deterministic and heuristic correlation techniques.
Q: Can you talk a little about the analysis capabilities Velocimetrics provides and how they are best used?
A: Velocimetrics’ analysis capabilities extend beyond system and network-oriented latency metrics, for example providing performance metrics as perceived by clients. As Velocimetrics captures business data as well as timing data, it can analyse flows at the business level, perhaps helping firms identify opportunities to reduce losses and increase profitability.
For instance, Velocimetrics can help firms diagnose if loss-making trades could be due to internal processing issues by enabling users to trace orders or trades through the entire execution process from origination to fill. This is possible even when an order has been split into multiple child orders or where a consistent ID fails to survive throughout the flow. Velocimetrics then enables the user to understand the latency experienced at each stage and evaluate if the performance of particular systems may be making an impact on trading success. A further example is how Velocimetrics can analyse trading patterns and alert users if certain clients are consistently trading outside a formal quoting period, but within the discretionary phase allocated to accommodate arbitrary technical delays. Velocimetrics can rank clients to help identify regular systemic abuse of this nature.
Q: What sort of firms is Velocimetrics designed for and how can they integrate it with existing systems?
A: Velocimetrics has been designed to be highly flexible and support a wide range of firm types and use cases requiring business flow monitoring. Velocimetrics is currently being used by exchanges, global investment banks, regional banks and spread betting companies. It is also being explored by hedge funds and extranet and financial data providers. Velocimetrics is able to integrate with many existing systems using a software development kit (SDK). It has been designed to work well with various types of applications, including Java, C#/.NET and C++ language bindings, an ever expanding library of middleware and exchange protocol decoders, and a decoder SDK to support analysis of proprietary protocols. As part of its integration capabilities, Velocimetrics can dovetail alerts into existing client systems and third-party enterprise monitoring systems, and consume information from these systems to enrich monitored data.
Q: What benefits does Velocimetrics deliver?
A: As Velocimetrics supports multiple firm types, trade lifecycle stages and asset classes, the benefits experienced by users differ. However, because Velocimetrics monitors entire environments, one of the primary benefits it offers to many firms is that when an alert is raised, users can instantly pinpoint the exact source of an issue, regardless of the environment’s complexity or whether the issue occurred on a network or within an application. This means large amounts of time and cost can be saved and preventative action can be instigated immediately, mitigating the issue’s potential impact on future orders, trades or clients.
Q: Can you support the market shift towards a balance of low latency and performance, rather than a focus purely on latency?
A: Yes, we can. Velocimetrics is already in use with a number of clients where insight into end-to-end operational integrity, rather than just latency, is the highest priority. The performance focused trading and processing strategies being instrumented are often very complex, with multiple moving parts which, if not effectively monitored and controlled, can generate significant operational risk.
Firms wishing to trade these strategies at ultra high speed need to be able to monitor the quality of the inputs as the potential for decisions to be made based on delayed, incomplete or unreliable data can quickly convert profitable trades into huge losses. Firms then need to ensure that these strategies are being executed as expected and that algorithms are able to retain the ability to make intelligent trading decisions in times of volatility. Velocimetrics enables firms to manage these factors by providing real-time insight into trading systems, without adversely impacting system performance. This provides users with confidence in the execution of low latency and high performance strategies.
Q: How do you expect the market to develop over the coming year and how can performance monitoring support these changes?
A: I expect we will see a number of developments in the coming year. As touched on earlier, trading speeds are starting to collide with the laws of physics and many firms that are keen to stay one step ahead of the competition are looking at how they can trade increasingly sophisticated strategies at speed. This trend is likely to continue in the coming year.
From a regulatory perspective we are seeing a growing emphasis on creating a more responsible financial system. Firms are increasingly expected to manage their systems’ compliance and integrity, which makes perfect sense, but doing so means they really need to be able to unravel the complexity of their systems and truly understand what is happening inside their trading environment. For many firms this is going to require a new approach. Pressures in the post-trade environment are also mounting as the level of tolerance for post-trade processing inefficiencies is in rapid decline and regulators are increasingly concentrating their efforts on this area.
These trends are likely to generate a wealth of challenges for market participants in the coming year. Effectively managing them comes back to the age-old adage that you can only manage what you can measure. Achieving a level of transparency that instills confidence will become an essential success factor as these trends evolve.