Why Latency is a Bad Bet
While training as an accountant, my father’s first job was at a bookmaker’s (betting shop to the uninitiated). On his first day, his boss gave him some very valuable advice.
“Son, don’t you ever bet. I don’t bet. The bookie (pointing to himself) always wins!”
Betting is big business in the United Kingdom. “Having a flutter on the horses” is an age-old tradition, and part and parcel of the British way of life.
You can take a bet on pretty much anything, just walk into the bookmaker and ask them to calculate the odds – when he was 14 years old, Chris Kirkland’s father placed a £100 bet at his local bookmaker that his son would one day play for the England Football team. 12 years later, Chris Kirkland came on as a substitute and his father won £10,000.
The Innovation Game
Things have moved on quite a bit since my dad’s day. Horse racing is still the number one sport bringing in bets, but betting on football is growing fast, along with several other sports. Technology is driving massive growth in the industry. In 2012, the total money bet online broke the £2 billion barrier for the first time, and is predicted to hit £3 billion in 2016.
In order to support this massive growth, the technology used by gaming companies have to be at the very edge of innovation. The main channels - retail outlets, web and mobile - all connect to infrastructure which is effectively a secure, low latency, high throughput electronic trading system. While most businesses are looking to reduce spending, in this particular industry, IT budgets have gone up "significantly" and so has the headcount. This trend looks set to continue.
The customer is king
Speak to any CIO in the industry, and there is one phrase that they all use again and again “smooth/good/reliable customer experience”. Generation Y has grown up with social media, and technology at their fingertips, and expect nothing but a reliable service when using their services on any channel. The platforms have to be updated at lightening speed, downtime is unthinkable, and performance has to be guaranteed, no matter what time of day. When you’re making an impulsive decision to place a bet, the last thing you want is for application performance issues to rear its ugly head.
Downtime or latency means loss of revenue, but it’s more than that. In this industry, your reputation and brand name are extremely important. A single incident or problem could drive your customers elsewhere, and that’s it.
One key problem is that demand is in no way constant. Just before any football game or race, there is a peak in the activity on all channels. But the busiest day in the year for any gaming company is without question – The Grand National.
The Grand National – It doesn’t get bigger than this
For the uninitiated, The Grand National is the most prestigious race in the UK (some say the world!), which started back in 1839. It is the most valuable jump race in Europe, with a prize fund of £1 million in 2015.
In 2015, nearly half of the UK population (including yours truly) placed a bet on the race, with total bets of around £300 million, with most of the bets placed on the day of the race itself. That’s a huge chunk income for a single day, but it also places massive demands on IT. The infrastructure has to be able to cope with this without a drop in performance. It’s a massive challenge.
Currently most companies accommodate for this by over provisioning their systems, and monitoring everything, reacting to any issues as they crop up over the entire weekend. Activity builds up throughout the day, and there is a spike just before the race. Once the race is over, there’s another peak in activity, as punters check to see if they’ve won anything.
Over the Grand National weekend this year (2015), I had the privilege to work with a betting company on a proof of concept. They put our Operations Manager to the test, letting it run and executing all our actionable recommendations over their busiest weekend in the year. The verdict was resounding, VMTurbo passed with flying colours - in real time, the recommendations were able to react to changes in application demand, and keep the environment in the desired state.
But don’t just take my word for it.
Latency – The enemy of profit
A study by Principled Technologies found that VMTurbo improved application performance through actionable recommendations—specifically cutting latency by 37%. After two test runs, the virtualized environment used in this study saw an increase in the number of orders per minute from 191.229 to 236,537.
Latency is a nasty beast:
- Amazon found that 100 ms cost them 1% in sales. In terms of its 2014 revenue of $88.94 billion, 1% is $889,400,000. Not insignificant.
- One study found that a 1 ms advantage can be worth $100,000,000 to a major brokerage firm.
- It’s so costly that $1.5 billion to lay the first trans-arctic ocean fiber optic cables in order to cut latency between London and Tokyo by 30% is a steal.
So we’re talking hard cash here. I’ll go out on a limb and say that VMTurbo can help you make more money!
So take a bet on us today, and download our 30 day trial and see how you can improve application performance across your infrastructure.