Sporting Group CTO on selecting Sumo Logic over Splunk, and Google over AWS
Peter Wallis explains how the company moved from Splunk to Sumo Logic because of pricing, and how Kubernetes made Google the clear cloud choice
Peter Wallis is currently chief technology officer (CTO) of Sporting Group, which has two divisions - a B2C operation called Sporting Index, which is a sports spread betting company, and Sporting Solutions, a B2B operation which supplies real-time pricing and sports trading capabilities to sportsbooks around the world.
His previous roles at Camelot UK Lotteries and ETX Capital mean that he has worked for nearly a decade in businesses that all have peaks in demand for services.
"I've been at companies where there are events, whether it's a lottery draw, a Millions roll over or the US presidential election when the pound sterling dropped several percent, events like that which have an impact on your systems, and suddenly you have a lot more data to handle - those are the challenges I enjoy," he says.
At Sporting Group, these events are, unsurprisingly centred around sports with a big focus on football. There's a huge amount of people betting on football games, and on a Saturday at 3pm, the amount of data that needs to be processed on the company's systems will multiple by two or three hundred per cent.
Switching from Splunk to Sumo
Sporting Group had been a Splunk customer between 2011 and 2017. Splunk's platform meant that Sporting Group had to pay for any peak usage.
"We had to pay for peak data ingestion volume, which we could potentially hit four times a month and the rest of the time we're paying for capacity we're not using, because what they do is they carve out part of their infrastructure for you and that includes the peak," Wallis explains.
Last year, the company opted to switch from Splunk to Sumo Logic, because its system runs on the cloud, so it can handle the peaks in volume without Sporting Group having to notify the company.
"With Sumo, they've got a much more multi-tenanted elastic proposition, so you can commercially pay for an average over a month on data ingestion rate, so that suited our business much more and we didn't feel like we were being punished for the nature of our sporting calendar driving peaky volumes," Wallis says, adding that his team prefers to have a partner that can scale on demand as it usually has four days a month where it needs to scale significantly.
"They just deal with it as we chuck it at them, their system scales and we don't see any degradation in our ability to alert on issues proactively, so these will all function as normal even though the data points have increased significantly," he says.
In addition, a lot of the features that Wallis and his team wanted to use with Splunk came at an additional cost, whereas with Sumo Logic this hasn't been as much of an issue.
"We asked Spunk ‘can we do this?', they said ‘yes, you can it's in this module which you don't have which would cost an additional amount', whereas with Sumo we don't have those conversations because so much more comes out of the box," he says.
The dynamic threshold feature for instance, is available out of the box with Sumo Logic but requires an additional charge for Splunk.
[Turn to next page]
Sporting Group CTO on selecting Sumo Logic over Splunk, and Google over AWS
Peter Wallis explains how the company moved from Splunk to Sumo Logic because of pricing, and how Kubernetes made Google the clear cloud choice
Moving to the cloud
Since the switch to Sumo Logic, Sporting Group has also looked at shifting its core backbone into the cloud into Google Cloud Platform.
"We do have a bit of AWS on the data storage side using S3 for financial reasons, but the reason we went for Google was because we were keen on Kubernetes, and as it was originally a Google product, their GKE (Google Kubernetes Engine) offering felt more mature than AWS or Azure, so felt like a safer option," he states.
Currently, the sports betting company runs its betting platform for its UK B2C platform on GKE in Google Compute, and it also has a risk management system which runs on Google, while it is rearchitecting a number of other systems to run in the cloud so that it can take advantage of what Wallis says is better scalability and pricing.
Sporting Group has also invested in some additional capabilities such as Cloudflare, which provides the company with DNS management services and distributed-denial-of-service (DDoS) protection. The company also runs MongoDB on-premise, as well as MongoDB's Atlas service which runs on GCP.
"Everything we look at now, we ask if we can buy a service to do something, because the team I inherited when I got here spent too much time patching servers and doing business-as-usual work, and what I want them to do is buy business value for things that they own and are responsible for," Wallis says.
"That means they can say ‘we don't need to run this database on-premise because we can buy a service that does this', and subsequently we won't need to worry about scalability or other factors - it means greater business agility and velocity to implement new things," he adds.
Wallis emphasises that this is not an infinite journey - it's a two to three year plan to shift everything onto GCP, but many of the other choices like MongoDB and Cloudflare are likely to be made on a continual basis.