Technology has changed quite a bit since the first Olympics event was held in Greece in 776 BC. These days, IT plays a critical role in the Games. Supporting the most watched athletic event on the planet requires solid systems that have undergone years of testing. Ensuring the athletes' performance is properly measured requires reliable systems with redundancy for all critical components.
To find out more about how the International Olympic Committee meets such stringent requirements, I caught up with Fred Wojciechowski. Head of Olympic Games Technology, Wojciechowski oversees technology programmes for the Winter Games with the next edition at PyeongChang in South Korea in 2018. He also manages contractual relationships with certain key partners for IT services for both the Summer and Winter Olympic Games.
How do you manage to deliver on time, every time?
Fred Wojciechowski: In most industries you can accept a delay - NASA even delays launching space shuttles. This is not the case for us. We cannot tolerate any schedule slippage. We know the date of the opening ceremony well in advance, and we simply have to deliver on time. We could never ask an athlete to run again because we could not capture his or her performance due to some technical issues.
When we begin planning for the Games, we start out by looking at the strategies and techniques that were used for the previous Games. Then we make adjustments based on multiple factors, such as new partners, the size of the organising committees, and new services. Finally, we add in a little buffer here and there to cover last-minute surprises.
One year before the games we have a major milestone, which is what we call the "test events", where we test the infrastructure for every sports venue, we test the people, and we test the processes. Those three things together - technology, people, and processes - make the success of what we do.
Based on what we find out during the test events, we have one year to adjust and fine tune. While we do have a nice buffer, in all the projects I've done, nobody sits around twiddling their thumbs. We stay very busy.
But the good thing is that people who do this job love what they do. They are passionate about the Olympics.
So you have no wiggle room on your schedule. What about quality? How do you ensure the delivery of quality systems?
Fred Wojciechowski: We have a very significant testing programme, which takes approximately 20 months for the Winter Games and 24 months for the Summer Games. Our IT integrator, Atos, perform a 100,000 hours of testing for the Winter Games and twice as much for the Summer Games.
We have a centralised test lab in Spain provided by Atos where all the systems are pre-tested prior to being delivered to the host country for the Games. We go through a long series of tests - stand-alone testing, interface, end-to-end, failover, disaster recovery, and anything else you can think of.
Of course we find mistakes during the testing. That's what testing is for. We even go into the Games knowing there are a few bugs, but really minor bugs that don't affect the SLAs [service level agreements]. We classify the defects into four categories, from most impactful (severity 1) to least impactful (severity 4). In order to minimise the risks associated with making a change at the last minute, sometimes we accept severity 4 defects, which are cosmetic defects that only we notice.
The two-year testing programme means our architecture and hardware platforms are frozen two years out. That's how we can be sure of having a stable testing platform and a stable system during the Games. While the hardware is frozen two years out, we do entertain software changes all the way up to two or three months before the Games.
We're also very careful about the maturity of the technology we use. It's always tempting to use the latest and greatest innovations. At the same time we want to remain conservative, because there's a certain risk in using the latest bells and whistles that haven't been tested as thoroughly.
That's not to say that we don't consider new technologies. We have specific committees that look at the different innovations that might be deployed. We group the systems into three categories - those that are critical, those that are necessary but less critical, and showcase systems, which may create a "wow!" effect. We are more accepting of innovation for the showcase apps than for the critical systems.
You must have enormous fluctuations in capacity requirements?
Fred Wojciechowski: That's right. Capacity is one of the things we have to plan well in advance. We perform detailed capacity planning for the entire life cycle of the project. For us, a technology project is five to six years long; and we break this down into multiple phases, with each phase carrying a different set of capacity requirements.
All the systems are analysed in isolation because each system requires different capacity management. During the games, the most important system in relation to capacity planning is the web site of the organising committee, where half of the planet will go to check for results.
Over the last few Games editions the capacity needs have been increasing at a very high rate. So we have recently made a decision to move to cloud computing, which will help us with the sizing and capacity of the different modules we need to host. During the most critical phase, we also use a content delivery network which allows anybody to access content anywhere in the world without delay.
One thing that's for sure is the fans are getting more and more enthusiastic. They want information all the time - and they want it in real time, whether from the website, social media, and so on. The demand is constantly increasing; and coping with that demand falls under the responsibility of those of us on the technical side.