Neueda Consulting is a Belfast headquartered software development & IT firm, and is emerging as a world class centre of excellence in delivering electronic trading projects, products and managed services to the capital markets industry.
Citihub Consulting’s Stuart English speaks to Malachy O’Neill, Head of Capital Markets and Colin Pattison, Head of Software Engineering, about inherent problems with testing trading applications, algo controls, disorderly markets and MiFID II.
Citihub Consulting: Tell us a little more about Neueda, the company history, key deliverables in the electronic trading space and current R&D initiatives?
Malachy: This year Neueda is enjoying its 10th anniversary. The company is a privately owned organisation with about 200 staff and associates- largely consisting of software engineers, business analysts and project managers. Recently, Neueda gained recognition for rapid growth, when shortlisted for the 2015 Deloitte Technology Fast 50 Awards. The company began life providing staff augmentation and IT transformation projects to the public sector in Ireland. But as the region evolved as a trading application development hub for several global capital markets institutions, Neueda’s business responded accordingly.
Today, we are working with global investment banks, exchanges, brokers, proprietary trading and market making firms, delivering custom software development projects, pertaining to market data, order routing, risk and algorithmic trading frameworks. We have also developed several licensed products, including a multi asset class, multi-protocol Exchange Simulator and a low latency middleware offering. We also deliver a cost efficient and cost predictable FIX customer onboarding service for major sell side institutions. Finally, R&D continues to be a strong focus for Neueda. Presently, we are evaluating the computing power of the Intel Xeon Phi co-processor and its potential application within options trading and risk analytics.
Citihub Consulting: Can you provide a description of the Exchange Simulator and its core functionality?
Colin: The Neueda Exchange Simulator attempts to replicate the behaviour of numerous exchange platforms and protocols, including native binary protocols and FIX, to provide a controlled environment in which clients can construct complex market scenarios and execute them in a repeatable manner. This can then be incorporated into development and testing processes to ensure and improve software quality.
The Simulator supports order entry and market data publishing across multiple asset classes. It can replicate different trading models e.g. scheduled and intraday auctions, and supports a selection of order types including but not limited to Imbalance, Iceberg, Pegged, Market and Limit orders. In addition to this functionality the framework has a number of Python extension points. These permit users to augment the message flow inline using the Python scripting language, and allows the creation of otherwise rare corner cases such as a broken trade message.
It has been designed to be lightweight and can easily sit within a developer’s virtual environment. Equally it can be deployed on high performance server infrastructure to complete stress testing.
Citihub Consulting: MiFID II requires conformance testing to ensure firms systems interact properly with the trading venue. Do you foresee the Exchange Simulator replacing this exercise?
Malachy: No, the Exchange Simulator is not a replacement for conformance testing with a venue. We believe it should be used in preparation for such conformance tests and can be used to ensure that conforming behaviour is maintained. It can also be used to support investment firm’s annual self assessment and validation reporting processes. Furthermore, if a significant material change to a trading system is planned then the Exchange Simulator can be leveraged throughout the development, testing and governance cycle to ensure the change process runs as smoothly and efficiently as possible.
Citihub Consulting: MiFID II places significant requirements on pre-deployment testing of algorithmic trading systems, as well as post deployment management. How does the Exchange Simulator help firms to meet these requirements?
Colin: Essentially, the Exchange Simulator can be used to ensure that algorithmic trading systems behave in an expected manner and thereby do not contribute to disorderly trading conditions. It can also be used to stress test systems during unusually high message and volume rates.
Customers can submit orders to the book to generate their desired message flow, or alternatively replay data into the Simulator. These features can be used to performance test algorithms to ensure that they meet the 2X trade data/message volume Microstructural Issues requirement, and more importantly it can be used to build extremely complex market events, beyond the minimum requirements, defined above. This can extend to altering individual parameters to the algorithm itself to test its behaviour under unexpected events.
Citihub Consulting: How would you propose an investment firm leverage the Exchange Simulator to improve their development/test processes?
Colin: When extending a trading system to add support for a new venue, a developer can leverage the Exchange Simulator in a test driven development process. Before developing the extension, a series of expected message flows for the venue can be created e.g. liquidity is added to the book programmatically and a new order sent that the developer knows will generate a partial fill. During development of the application the test suite will grow and can be regularly rerun to ensure that behaviour is maintained and there is less regression.
Once the development process is complete venue conformance testing can take place. On successful completion any gaps in the test suite can be addressed and coverage improved. The test suite should be incorporated into nightly integration processes within the continuous integration server of choice. For future development cycles any required conformance test should be straightforward, as all commits have been tested for conformance against the existing test cases.
In order to test behaviour of an algorithmic trading system a dataset(s) would be designed with parameters carefully designed to test the algorithm in unusual market events for example large price deviations. This design process would involve the investment firm itself, and any third-party involved in the development of the algorithm itself to ensure a strong working knowledge of the design.
Citihub Consulting: How do customers access the Exchange Simulator?
Malachy: The Exchange Simulator can be deployed on premise or customers can connect to the cloud hosted platform. It’s available on a free trial basis, so customers can evaluate the benefits of using the Exchange Simulator as an integrated component of their development, testing, and governance cycle, prior to commitment.