Running after the Olympics

In 2004 the already financially troubled Greek country hosted the Olympics. It needs no explanation that organising one of the biggest sporting events on the planet comes with a huge investment. For one thing, you need to build the sport venues, which are of course the main scenery of the event. Less visibly, the supporting infrastructure like hotels and transportation, is proving to be a massive challenge for every host of such happenings. For the Athens 2004 Olympics, the building phase turned out to be a struggle. And though some of the facilities were finished only just in time, the 2004 Games are generally seen as a successful edition. The bigger problem, however, started after the Olympics. Most venues are abandoned in the meantime and have never seen any use since the end of the games. Building them was one thing.  Running them turned out to be a much bigger challenge.

The 2004 Olympics show that Athens was not built in one day. This sounds very much like big IT projects. The build is a challenge, but instead of focusing exclusively on a stable arena, you need to focus on how you want to sustainably exploit it in regard to the future.

Sounds familiar? You made plans for a new IT platform or new business application.  The deadline has been met just in time. Next, everything is put into production and a celebration is well deserved. In traditional, on-premises scenarios, the run (or operate) is usually planned and ready for execution as well.  After all, the environment is stable, the ITIL processes are robust and we’re talking about a planned number of deployments in the near future.

Enter cloud.

With the introduction of the cloud, your IT vision changes. The new idea is to have regular deployments in much smaller parts.  Continuously. The idea is also that the ones building the applications are the ones who will run them as well.  And that is exactly the pain point of the story.  In large, complex enterprise environments we often see an Athens 2004 scenario.  To the point where a major critical situation happens and the “you build it, you run it” principle fails.

Let’s have a look at what you can do to avoid this situation.

Enterprise environments are rarely built in a cloud only setup.  There is a cloud component and there is an on-premises one.  More and more, we are confronted with multiple cloud components from different clouds.  Which is also called a hybrid setup.  While this reality is offering an agile architecture to create business value, it comes with significant challenges for the enterprise scale operations. What do you need to do to validate your service management processes for a hybrid world?

1. Never outsource a problem

First things first. We regularly run into the scenario where companies initially want to organise the cloud operations themselves, rather than outsourcing it or look for a managed service. As the process matures, they face the realisation that this might not be a real viable option. As a result, the balance swings to the other side and the search for a managed services provider starts. There is something to be said for both scenarios and also for a combination of both. But one thing is not negotiable: you must first have a very good understanding of your cloud or hybrid platform from a technology and a governance point of view, before you bring a managed services provider to the table. If the current state of the platform is problematic, nothing will magically be fixed by bringing in an outsourcer.

2. Set up a realistic accountability model

Secondly, the famous accountability discussion does not get any simpler in a hybrid context. It is not a black and white story. We have seen scenarios where the strategy was to get full autonomous product teams in place, without defining what autonomous exactly means. Safe to say: this is not going to work. Even worse, how autonomous looks like, is affected by architectural decisions. A modern cloud application is a very different scenario from – for instance – an IaaS based virtual machine lift and shift scenario. Not to mention the financial implications of each scenario for the different teams involved.

3. Monitor and remediate

At the end of the day, it does not really matter who is responsible for which task and how the organisational design is set up. The effort that needs to be delivered to keep an enterprise platform and the associated workloads running is rather fixed. Monitoring that each task is not only owned by somebody, but also actually picked up is key to avoid failures with significant business impact. If monitoring-hooks are not embedded in the code, it does not matter how brilliant the operational dashboards are. They will be empty.

In conclusion, the parallels drawn between the challenges faced by the Athens 2004 Olympics and large-scale IT projects, particularly in the context of cloud adoption and hybrid environments, reveal interesting insights.

The narrative underscores the complexity of transitioning from traditional, on-premises setups to continuous, cloud-based deployments. As shown, the 2004 Olympics were a single use set-up. Meaning the “you build it, you run it” principle, while ideal in theory, meets significant hurdles in enterprise environments, similar to the post-Olympics scenario in Athens. The hybrid reality, with a mix of on-premises and multiple cloud components, presents both opportunities and challenges for agile architecture.

---