Cryptocurrency values can rise or fall fast, so trading depends on liquidity—the ability to buy or sell quickly before prices change. GSR, founded in 2013, has a global footprint and provides cryptocurrency token issuers, exchanges, financial institutions, and investors with that liquidity.

During the COVID-19 pandemic, GSR saw a rapid rise in demand for cryptocurrency trading, as many governments created large financial stimulus programs and large investors put more money into digital assets. To improve its systems’ speed and scalability, it turned to Amazon Web Services (AWS).

Using AWS, GSR has added support for more than 1,400 trading instruments, which let customers conduct transactions in many different currency combinations. It also gained the elasticity it needed to serve its growing customer base, helping it to increase the size of its workforce by a factor of 5. And it’s gained the ability to manage daily trading volumes that, at times, have reached values of over $5 billion.

GSR has roots in traditional finance, with many of its executive team members coming from the likes of Goldman Sachs, Citadel, and Two Sigma. In addition to providing liquidity services, it also manages trade derivatives, supports over-the-counter trading, and creates custom-made trading algorithms. 

The company needed to update its IT systems quickly to meet a rise in demand for cryptocurrency trading during the pandemic. It had already been using AWS and it looked for opportunities to expand its use. 

The goal was to ensure scalability and fast network connections. “There was a lot of trading happening, and a lot of new liquidity coming into these markets,” says Matteo Cerutti, head of trading platform at GSR. “There was just this massive influx of interest into the sector. I think that started this next wave of the market.” 

Building on GSR’s existing foundation on AWS, the provider was a natural choice. “AWS is very reliable,” Cerruti says. “I think it’s very hard to beat.”

GSR launched several projects to improve its infrastructure with the help of dedicated AWS account managers and technical teams. 

The addition of three new AWS Availability Zones allows GSR to provide faster regional connections to exchanges. “For trading at very high speed, that kind of connectivity is very useful,” says Cerutti. For services in the New York area, GSR uses an AWS Local Zone, which places AWS compute, storage, and database services close to large population centers. This means that GSR can run applications with single-digit-millisecond latency. 

GSR also optimizes costs by working with AWS solutions architects to use AWS Reserved Instances, which provides discounts of up to 75 percent compared to buying capacity on demand. 

Automated trading models help GSR to rapidly manage transactions, enabling it to handle more than 1.1 million trades a day and, at times, over 100 million daily orders. GSR supports these models by using Amazon Simple Storage Service (Amazon S3), an object storage service that lets it retrieve any amount of data from anywhere, and ato run batch computing jobs at any scale. It manages data using Amazon Aurora, which is designed for unparalleled high performance and availability at global scale with full MySQL and PostgreSQL compatibility. 

Using these services, GSR’s research team can access the data it needs to analyze trading results. The trading team then develops automated strategies to monetize trading signals. “The main trading that we do is done programmatically on exchanges,” says Cerutti. “So, if you go to a crypto trading platform today, and you see the order book going 100 miles an hour, and then you put in your bid for Bitcoin or Ethereum or another crypto asset, we’re there to sell it to you and we’re also there to buy it from you. We’re both sides of that order book. And we’re doing that at scale.”