Sporting Solutions chooses MongoDB for scalable odds-pricing engine

Data provider Sports Solutions has deployed MongoDB NoSQL tools to power its in-play betting engine for customers such as William Hill, providing the scalability that could not have been achieved with traditional database technologies.

Share

Data provider Sports Solutions has deployed MongoDB NoSQL tools to power its in-play betting engine for customers such as William Hill, providing the scalability that could not have been achieved with traditional database technologies.

The British firm is a subsidiary of Sporting Index, which has provided services to consumer customers for 22 years. As part of the firm's move into the B2B market, Sporting Solutions launched its Connect platform in 2012 to provide pricing to other betting firms based on its proprietary algorithms.

In order to build a pricing engine platform capable of supplying the greater volumes of data required by its customers, the firm decided against using relational database technologies and looked to MongoDB’s NoSQL software. This provided a number of benefits such as the flexible schema, speed to deploy and ability to scale.

“It would have been challenging from a scalability perspective [to build Connect on SQL], as you could use a SQL database in a NoSQL fashion but then you are doing a lot of it yourself,” said Andy Britcliffe, CTO at Sporting Solutions.

“A lot of the things you get then around MongoDB – portioning capability, horizontal scalability – mean that while you could do it, it wouldn’t be advisable.”

Sporting Solutions also considered other NoSQL tools in what is an increasingly competitive market. MongoDB, which received $150 million funding last October from the likes of Intel and Salesforce, came out on top compared to others including Couchbase.

“We evaluated a number of other products in the space. There were three key factors we were looking for – the developer productivity tools, operational effectiveness, and thirdly the eco-system around it. MongoDB came out strongly on all points.”

He added: “Speed was a key factor - our data is ephemeral but needs to be delivered in a real-time manner. When we talk about in-play updates of football, every second counts. So for writing and reading data, MongoDB was extremely fast.”

The engine is built on an OpenStack platform, with its 400 virtual machines split approximately 50-50 across its on-premise data centre and the Rackspace public cloud.

The Connect platform also feeds into Amazon’s Hadoop platform, Amazon Elastic Map Reduce, and the Amazon Redshift data warehouse for analysis purposes, with data stored in AWS S3 storage ‘buckets’.

Going forward, Sporting Solutions expects to expand its use of MongoDB more widely across the business. This means the potential to rework the betting engine at the heart of its B2C operations to run on MongoDB, rather than the legacy technologies it currently requires.

For example, this would allow the company to expand the services it offers direct to customers, such as providing extra fixtures and extra markets.

“The pain-point with the legacy technology on the B2C side is that it doesn’t scale, and is more built around old style monolithic applications,” Britcliffe said.

“What we have seen with building a distributed system powered by scalable NoSQL technology is that it gives a lot of headroom for scale, and that is what we want to achieve for spread betting.”

Find your next job with computerworld UK jobs