Amadeus is critical to the travel industry. The global distribution system (GDS) for airline tickets was established...
in 1987 by Air France, Iberia, Lufthansa and SAS to provide services to connect travel agents with airlines.
For Wolfgang Krips, head of global operations at Amadeus, automation and a highly skilled team work in unison to provide high-availability IT services. Krips has worked at Amadeus since July 2013. Prior to joining he was managing director of SAP managed services.
As head of global operations, Krips is responsible for ensuring the services Amadeus provides are available 24/7, 365 days of the year.
Krips is based at Amadeus’s Erding datacentre outside Munich, which, prior to the likes of Google, Microsoft and Facebook, would have been regarded as one of the biggest European datacentres. Today it operates the largest datacentre in Europe for the travel industry.
"Our systems build up travel itineraries and book your travel," he says. This is the core Amadeus global distribution system that enables travel agents to check seat availability and book tickets for passengers.
Since 2000, the company has been expanding to offer a number of cloud-based products.
"We also support the operational processes of airlines, such as check-in, printing baggage tags, boarding and the load plans for the aircraft," says Krips.
More on high availability IT operations
These systems are mission critical to the airlines. "If these services are not available, the aeroplanes are unable to leave the airport," he says.
Krips and the operations team at Amadeus are responsible for keeping these systems available at all times. Any failure could lead to major travel disruptions globally. "If we have a system outage, it affects the operations of the airlines," he says.
The site in Erding is made up of three separate datacentres, each operated in an N+1 configuration which provides additional hardware for failover.
“Everything in the datacentre is redundant. We can operate for days even if we were cut off from power and water. We have redundant computers and state-of-the-art failover technology,” says Krips.
Amadeus makes extensive use of automation.
"An automated environment is one of the ways we fulfil the requirements of the business. No matter how fast a human works, it can take a week or so to bring a physical server into a datacentre," says Krips.
"We now live in an age where you can go to Amazon in the cloud to get a virtual server instantly, so the classical approach of physically provisioning a server needs to be decoupled using virtualisation and automation," he says.
But Krips says datacentre automation is impossible if there are too many configurations, in terms of the hardware and software set-up. "We are trying to standardise the technology to reduce complexity by having as few variants as possible. From here you can automate," he says.
Operations has moved from box-shifting to a very sophisticated discipline where operations staff have to think through how to make processes fail-safe
Wolfgang Krips, Amadeus
Fewer datacentre configurations reduces complexity and increases agility, according to Krips. "If you need to bring in automation that covers 85 variations of hardware, it takes much longer to deploy and test. If you have only two or three variations of hardware, complexity is less and the configurations can be automated more easily," he says.
He says there needs to be a balance between having the highly skilled team who can solve complex issues that arise when running a datacentre like the Amadeus site in Erding, along with the procedures needed to repeatedly get high quality results. This procedural approach to change management fits alongside the automation tools: "The human factor is very important".
Krips says standardisation plays a major role in reducing errors within a datacentre.
"Operations has moved from box-shifting to a very sophisticated discipline where operations staff have to think through how to make processes fail-safe. We have highly educated and skilled people who understand the systems they operate," he says.
The team adheres to a process framework to reduce the risk of human error. Krips says the operations team uses checklists, analogous to the pre-flight checks run on the flight deck of an aircraft, to minimise human error.
Career progression in IT operations
In many ways, Krips regards automation as a way to run the mundane tasks required to operate a major datacentre, albeit extremely quickly, allowing the highly skilled operators to problem-solve situations that automation cannot handle.
"Our datacentre services are automated to minimise any service disruptions. But automation means the skills required by operations staff increase. They no longer think about how to do the job, but instead they are asking, 'What is the next level of automation I can use?'," he says.
Due to datacentre automation, he says a very different level of thinking is required.
Speaking about the progress of IT and the effect on IT operations, he says: “In the early days you could bring in any technology and the business would figure out what to do with it. But today you need a business case.
"You need people who not only understand the technology, but also understand the price in terms of operational costs and the total cost of ownership. This is what is needed for the managerial side of IT operations.”