In a football-crazy country with 200 million people, and a truly global customer-base beyond that, handling 1 million unique visitors in a single day is a special challenge.
But an even bigger challenge is the fact that all of those visitors ultimately want a ticket for the FIFA World Cup™ final. In order to deliver performance, reliability and scalability whilst at the same time managing end-customer expectations, MATCH Services have implemented leading edge cloud technologies for the 2014 FIFA World Cup Brazil™.
Taking us through how their leading edge cloud technologies are being utilised are –
- Tony Pearson, Business Consultant FIFA World Cup™, MATCH Services, Spain
- Chris Budd, Project Director, MATCH Serviços de Eventos, Spain
Prior to their presentation in March, Tony and Chris provided some key insight into the principles and specific analysis they considered before moving to the Cloud –
1. We took our time in choosing the right partners to fit our business model…
We didn’t only consider the hosting of the applications and data, but we also needed to consider caching and queuing technologies. We included in our analysis the services offered for backups and disaster recovery. We needed to ensure PCI compliance and to consider particular security requirements. We also had specific data protection legislation to adhere to. Finally, we also paid close attention to the maximum transaction throughput we would attain.
2. We didn’t fall into the trap of thinking that it’s all or nothing…
We decided that we didn’t have to migrate everything at once. If you’ll have to rewrite a lot of code just to enable the transition, then maybe you are being too ambitious.
3. We did some maths…
What were our financial objectives of moving to the cloud? We weren’t going to sell more tickets. We didn’t have any other event to cross sell. Then the objective was purely to reduce hosting costs (and we have). But then we forgot to include everything in our costings. If you go through the same don’t forget, for example, to include the costs for your research, setup, learning curve, development and (especially) testing activity, because you’ll get a bill for those too.
4. We did some more maths….
Our ticket inventory and sales model were fixed. We had a lot of tickets but not nearly enough to satisfy demand. We had unpredictable (huge) demand peaks and wanted to reduce the possibility of our sites being swamped. The decision we had to take from an economy of scale perspective was, “is there any real point in letting more customers into our shop than we have tickets to sell?” Did we want to focus the solution on serving promptly the customers we could serve, rather than allowing millions of customers, who we could not possibly serve, wander around our site?
5. We did even more maths….
Would our payment gateway be able to handle the throughput? Would our other third parties increase the total transaction time or become a bottleneck? How many single points of failure would we still have even after migrating to the cloud and how many of those would be still out of our control?”