Trading system functionality made possible by the cloud – Insights from Trading Technologies
Trading Technologies (TT) began its journey in the cloud more than five years ago. On the AWS Blog in 2015, we highlighted their cloud migration in terms of platform transformation and the cultural shifts they and their clients made for a successful conversion.
What has TT been up to since that time?
“In 2015, we shared our vision, and now in 2020 you’re seeing the results,” says Jason Shaffer, TT’s Chief Product Officer. As of October 2020, 80% of the migration to the new TT platform has been completed — with December as their target completion date.
Looking back to the vision TT’s CEO Rick Lane shared in the 2015 blog posts, Jason reinforces why they chose the cloud, rather than creating the next-gen version of their legacy trading platform: “The idea was not to go through this massive investment upgrade of technology where we’re simply giving a better trading screen.” He continues, “It was really so we could take advantage of the added benefits that cloud deployment offers, while allowing us to grow as a company and a technology provider over the next ten years.”
Their hybrid model only keeps a minimal footprint on premises to address particular latency concerns. Jason shares, “We use the data center for all the low-latency order routing and mission-critical, latency-sensitive operations. We put pretty much everything else in the cloud.” This enables them to take advantage of all the best cloud features and functionality, including security, scalability, the elasticity to handle market volatility, and compute on demand, while meeting customer needs.
Jason shares, “The growth we see in our products and services is much more in the data-cloud space than in having a physical presence.” Their legacy futures platform, notes Jason, “was really intended to be a closed superhighway between a trader and the exchange”—but the cloud opened up TT’s platform to partnerships, integrations, and wider data distribution. Expansions in their service offerings include:
- OMS: Inbound demand from traders drove the creation of an order management system (OMS). Having access to the cloud gave the two TT engineers on the project room to experiment. After a few days, they created a simple prototype. Jason says the prototype “made it real in potential customers’ eyes.” TT’s screen allowed them to fold their existing execution management system (EMS) and OMS onto one screen. The ability to do real-time alerting and an audit trail of those alerts is unique to TT’s OMS system, and is possible because of the cloud’s cost-effective data-storage capability.
One significant win was with the Institutional Services division (IS) of TP ICAP, which was another confirmation of their CEO Rick Lane’s cloud vision. “It was really a very high-level conversation about where they saw their business going over the next ten years and the changes they knew that they needed to make to evolve and stay profitable and stay relevant in the shifting tides,” Jason recalls. “And they saw TT as the strategic partner that they needed. And, so, really, the OMS deal was bigger than just OMS. It was an investment in what we’re doing as a company.”
Dan Feldman, TT’s VP of Systems & Networking Engineering adds of TT’s OMS launch, “When I learned that I didn’t have to build another co-location center, and I didn’t have to rebuild an entire hybrid cloud architecture for VPC direct connect mesh, I was the happiest person you could imagine. We were able to use the cloud to bring a new product feature to market in a fraction of the time leveraging our existing global footprint.”
- TT Score: TT Score, a surveillance and audit solution, uses pattern recognition based on machine learning (ML) to identify trading behavior that could pose a regulatory or compliance risk and uses ML training data and models derived from actual regulatory cases. “It’s one of the few ML-based surveillance tools on the market,” notes Jason. He adds that others are “rules based, manual, and labor intensive, generating lots of false positives.”
In addition, customers don’t need to trade on the TT platform to use this tool. They can simply send a FIX session to the TT platform to run it through a live monitoring environment. During the pandemic, traders have been working from home. “This too gives risk administrators the ability to see the behaviors of traders and nip any potential problem in the bud,” Dan notes.
- Backtesting: To let customers more efficiently develop trading algorithms, TT engineers built a simulated matching environment that pulls in historic market data as it happened on the day it was recorded for backtesting models. A customer can use the model to build an algorithmic trade strategy with an automated process that sends the model to compliance for review. When approved, the strategy can go right into production.
This simulation environment is offered as an Amazon Machine Image (AMI), allowing customers to deploy the UI and backtesting environment with a few clicks. When initiating the backtest, the UI calls a REST API backend that runs on load balanced Amazon EC2 instances, which utilize Amazon RDS and Amazon ElastiCache. The backend launches an EC2 instance based off of the AMI. Upon completion, the results are copied to an Amazon S3 bucket for long-term storage. Once usage statistics are written to the billing database, the EC2 instance is terminated.
Those solutions are just a sampling of how TT utilizes the cloud. Other AWS services in use include: Amazon Keyspaces (for Apache Cassandra) as a single source of uneditable truth for trade history for each client, where every execution-related message is encrypted, stored, and able to be retrieved—leveraging the best of what the cloud offers when it comes to scalability and cost. To manage a hybrid environment, they work with AWS Elastic Beanstalk, moving to a mix of containers and AWS Lambda functions. While they still have a few on-premises, ultra-low-latency deployments, that footprint continues to shrink as more systems can migrate to the local availability zones—along with further AWS enhancements to address the latency hurdle, such as AWS Outposts and the ability to run multicast in the cloud, using AWS Transit Gateway.
Before deciding to join TT, Jason was a trader and user of the legacy TT platform. “I was the first customer trading on the TT [hybrid] platform in early 2014 risking real money.” He says, transparently, of Rick’s hybrid cloud vision at the time, that he “was skeptical of what he was trying to do.” However, as Jason quickly saw the benefits of the new TT Platform, he bought into Rick’s vision. Towards the end of 2014 he said to Rick: “Listen, I get it now. I’m all in. Let me know how I can help. That conversation led to my current role I have here.”
Interested in deploying your trading system in the cloud or learning more about Trading Technologies? Get in touch!
About Trading Technologies
Trading Technologies (www.tradingtechnologies.com, @Trading_Tech) creates professional trading software, infrastructure, and data solutions for a wide variety of users, including proprietary traders, brokers, money managers, CTAs, hedge funds, commercial hedgers, and risk managers. In addition to providing access to international exchanges and liquidity venues via its TT® trading platform, TT offers domain-specific technology for cryptocurrency trading and machine learning tools for trade surveillance.