You are here:
Profile
Dr. Hans-Joachim Popp: who holds a PhD in electrical engineering, has been CIO of the German Aerospace Center since 2005. His previous posts include Product Manager at Dräger Medizininformatik, CIO at TÜV Product Service and IT Operations Manager at TÜV Süddeutschland.
Quotation
“Our success is gauged by how quickly and how well we can meet our customers’ wishes, even seemingly impossible ones.“
Dr. Hans-Joachim Popp,
CIO of the German Aerospace Center (DLR)
“We benefit from the space-rocket bonus.”
Dr. Hans-Joachim Popp, CIO of the German Aerospace Center (DLR), talks with Ulrich Strasse, Key Account Manager for DLR at T-Systems Solutions for Research (SfR) about standardization and flexibility, collaboration, transparent service delivery, and the competitive drawbacks of what he dubs the “Bayern-Munich effect.”
Dr. Popp, there is a natural conflict between the need for standardization of end-user environments and the need for customization. How can we resolve this dilemma?
Yes, that’s a tough one, but things are looking up. Today’s PCs are so powerful that they are ideal for most research work. If you compare the office occupied by a solar scientist with that of an accountant, you generally won’t see any real difference in the computers they’re using. But standardization has its limits. My team is constantly trying to determine what those limits are. An IT service provider needs to make sure that its standardized services don’t impair productivity, and at the same time, that economies of scale are fully exploited.
Yes, that’s a tough one, but things are looking up. Today’s PCs are so powerful that they are ideal for most research work. If you compare the office occupied by a solar scientist with that of an accountant, you generally won’t see any real difference in the computers they’re using. But standardization has its limits. My team is constantly trying to determine what those limits are. An IT service provider needs to make sure that its standardized services don’t impair productivity, and at the same time, that economies of scale are fully exploited.
How do you find a happy medium?
We try to gain a deep understanding of our researchers’ needs and separate the technical issues from factors such as habit and personal preference. I often compare it to an optical bench made of granite. To produce sound results, the table must make conditions as stable and consistent as possible. The scientist has to be able to concentrate on the experiment and not be hampered by bad tools. And it’s the job of the IT provider to deliver this kind of complete reliability.
We try to gain a deep understanding of our researchers’ needs and separate the technical issues from factors such as habit and personal preference. I often compare it to an optical bench made of granite. To produce sound results, the table must make conditions as stable and consistent as possible. The scientist has to be able to concentrate on the experiment and not be hampered by bad tools. And it’s the job of the IT provider to deliver this kind of complete reliability.
So that means IT innovations must be thoroughly tried and tested before being used in a research setting?
Exactly. We’re naturally extremely interested in IT innovations, but there is no place for them in experiments unless they relate to IT. We like to eliminate risks and uncertainties wherever possible. Many of our experiments take the form of computer simulations performed on systems that process data very quickly and very accurately. The worst thing that can happen is for a high-end computer to crash after two weeks’ work, causing the loss of valuable results. The fastest computer is of no use if it keeps failing.
Exactly. We’re naturally extremely interested in IT innovations, but there is no place for them in experiments unless they relate to IT. We like to eliminate risks and uncertainties wherever possible. Many of our experiments take the form of computer simulations performed on systems that process data very quickly and very accurately. The worst thing that can happen is for a high-end computer to crash after two weeks’ work, causing the loss of valuable results. The fastest computer is of no use if it keeps failing.
But when it came to outsourcing your computing resources, you were willing to experiment. Why is that?
That’s not true. In the past, researchers insisted on keeping their computers close at hand, literally under their desks – for emotional reasons, and because there was a lot of insecurity in those days. But it’s essential to locate our systems where they can be best managed – in terms of critical factors like the power supply, air conditioning and physical access control. That’s why T-Systems initially consolidated the equipment at local data centers, and now we are gradually moving them to a central facility. So outsourcing to T-Systems was not really an experiment, quite the opposite. It was based on a cautious, low-risk strategy; although admittedly, we were the first in the scientific community to take this route.
And there’s another factor to consider: high-performance computers are developing so fast that new hardware has to go into operation without delay. When an institution is responsible for its own IT, this can cause problems. That’s why I said: SfR has a ready-made, proven server architecture. It may not be the hottest technology around, but the main thing is, it will be up and running straight away. And we were able to migrate more than 20 clusters to a standardized environment operated by T- Systems.
That’s not true. In the past, researchers insisted on keeping their computers close at hand, literally under their desks – for emotional reasons, and because there was a lot of insecurity in those days. But it’s essential to locate our systems where they can be best managed – in terms of critical factors like the power supply, air conditioning and physical access control. That’s why T-Systems initially consolidated the equipment at local data centers, and now we are gradually moving them to a central facility. So outsourcing to T-Systems was not really an experiment, quite the opposite. It was based on a cautious, low-risk strategy; although admittedly, we were the first in the scientific community to take this route.
And there’s another factor to consider: high-performance computers are developing so fast that new hardware has to go into operation without delay. When an institution is responsible for its own IT, this can cause problems. That’s why I said: SfR has a ready-made, proven server architecture. It may not be the hottest technology around, but the main thing is, it will be up and running straight away. And we were able to migrate more than 20 clusters to a standardized environment operated by T- Systems.
