Opinion: Website performance - don't back the wrong horse

Inability to supply enough bets to meet customer demand during traffic spikes is, frankly, inexcusable. All the technology to prevent this happening is out there – tried and tested. And yet, only last January the National Lottery website crashed as people scrambled to place bets for the ?54 million jackpot.

None of this bodes well for the Grand National. Traffic spikes and website outages on this day are becoming as legendary as Red Rum. Before the 2012 Grand National, Paddy Power gave out free bets to its customers after its website crashed for 20 minutes. Last year, many bookmakers did the same.

Sites don’t have to scale their servers to deal with the spikes. Thoroughbred websites like Pinterest and Vimeo, with high bandwidth requirements and massive, event-driven spikes, use a new class of software in their web architectures to handle this.

The software doesn’t fit into a clearly defined category – yet. The IT analyst firm Gartner referred to it as “Web Scale Platforms” in its “Cool Vendor Report” last year. We call it “Web Architecture Performance”. What’s important is how these tools can give a serious competitive advantage.

Here’s a layman’s guide to the top five Web Architecture Performance concepts, so you can talk to your techies about how they might improve your website:

1. Caching (aka http caching, reverse proxy) – the process of storing data in a temporary location. The files you automatically request when viewing a Web page are normally stored on a server in your browser’s directory. When you revisit a recently viewed page, the browser can get those files from the cache instead of the server, saving time and network traffic. As such, caching plays a fundamental role in website uptime on days like Grand National.

2. Cache invalidation – related to caching, this complex process involves deleting cache entries and essential for betting websites to have set up optimally. If a web page cache isn’t cleared when new data is available, users get a stale version of the page – a crisis for a betting website that needs to deliver a fair, level user experience.

3. Load balancing – this involves divvying up workload between two or more computers to raise output and serve users faster. High-traffic betting and gaming sites usually use load balancing. It’s a key defence against the menace known as the DoS (Denial of Service) attack, anonymous hackers use to take down betting websites.

4. Application Programming Interface (API) management – An API is computer code that lets two software programs talk to each other. API management involves publishing, promoting, overseeing and creating resources for APIs. Betfair’s Exchange API, for example, gives developers access to market data, including pricing and volume.

5. Content Delivery Network (CDN) – this is a distributed network of servers, delivering users web content from the servers nearest to them. In a CDN, multiple copies of the same web content are mirrored on various networked servers. People accessing websites get content served up quickly, even when bandwidth is limited or there are sudden spikes in demand. Many betting sites already have CDN infrastructures, but can get much better performance during traffic spikes when used together with caching software.

So instead of betting on hardware to avoid a crisis, websites that want to maximise their transaction fees on big days like Grand National should learn from the industry thoroughbreds and invest in modern web architecture performance technology.

Go on, have a flutter!