Bet365 improves latency and peering by updating network infrastructure

Online gambling site Bet365 upgraded its network and improved its latency to deliver a fast, smooth and reliable experience to its customers

For the online sports betting service, Bet365, a fast and reliant network infrastructure is core to its IT architecture. The company upgraded its network to meet this need.

The online gambling website wanted to offer its customers the best “in-play” experience possible. 

Whilst competitors focused on replicating the offline experience online, Bet365’s founder Denise Coates decided to focus on the technically challenging in-play betting – the ability to make a bet while a sport is in progress, according to Bet365’s IT team.

The company employs over 2,000 people, is live in 17 languages and aims to deliver an uninterrupted and fast online experience to over 10 million customers.

But offering best in-play experience means the company’s IT team must ensure a robust network infrastructure to make services instantly available and to ensure that a continuous stream of up-to-date odds information is delivered to its customers in near real-time.

“Bet365 is a latency-sensitive business. To get the best from our platforms, we need a network provider whose peering connections can enable us to deliver odds information in near real-time,” says Neil Selby, head of networks and security at Bet365.

Peering is the arrangement of traffic exchanged between Internet service providers (ISPs). Larger ISPs with their own backbone networks agree to allow traffic from other large ISPs in exchange for traffic on their backbones.

But Bet365’s old network infrastructure, which was provided by a tier 2 supplier, was proving unreliable and incapable of coping with pressure of real-time betting. The IT team decided it needed a global carrier rather than multiple regionally strong network providers.

With a global tier 1 provider, the IT department implemented an efficient and resilient global IP network design, which will deliver fast user-experience and optimum geo-routing for traffic.

“We recognised the need to work with a true tier 1 provider to enhance our performance, resilience, and security,” says Selby.

The company then selected global network provider NTT and implemented a new network infrastructure in its London office.

NTT Europe now provides Bet365 with 10G feeds from its Global IP Network across Europe into each of its main hosting sites based in London and Manchester.

Before selecting the network provider, Bet365’s IT team used Keynote, an independent internet performance monitoring tool to simulate user experiences globally.

The results demonstrated that the new network infrastructure can dramatically improve traffic routing efficiency. In the simulated user test, Bet365 also set up IP anycast and proved that access was significantly more reliable.

In Internet Protocol (IPv6), anycast is a communication between a single sender and the nearest of several receivers in a group. It is designed to let one host initiate the efficient updating of router tables for a group of hosts.

A major benefit of using IP anycast is that it will allow bet365 to operate through a single IP address, according to NTT Europe. In the event of a regional platform going down, customers are automatically rerouted to the nearest site without further intervention. This strategy offers strong back-up resilience and disaster recovery options, the network provider adds.

Deploying IP anycast onto NTT’s network enabled natural geo-routing to the nearest serving platform, and strong peering points meant low-latency for additional performance. For Bet365’s customers this resulted in a fast, smooth and reliable experience.

But the IT team did not stop there. It recently added DDoS Protection Service to compliment the anycast architecture for security and to continue services at the time of a threat.

The new network and peering infrastructure from a single global provider improved geographic routing and reduced delays, Selby says. “Using NTT’s network globally, we went from 84 % accuracy to 96% accuracy of customers hitting the nearest POP [Point of presence],” he adds.

Read more on Datacentre performance troubleshooting, monitoring and optimisation

Data Center
Data Management