"Things are moving fast at the DFN”
Photo: DFN-Verein
Congratulations, DFN: The association for the promotion of a German research network operates an efficient fibre-optic network and the X-WiN scientific network, which researchers use to exchange data, communicate and organise their work.
A 10,250-kilometre-long fibre-optic network is the base for the X-WiN science network. This connects 849 universities and research organisations, it enables many IT services and strategies for IT security, in particular it faciliate the communication and the sharing of data between researchers. The association for the promotion of a German research network, known under the German acronym DFN, celebrates now its 40th anniversary. Its history is also one of technical progress and successful self-governance in science and research. More than 350 members work together to further develop the X-WiN and the IT services of DFN association, even though they often have very different needs. The international network offers fascinating opportunities and provides researchers and students with practical examples of how digital independence can be achieved between vast amounts of commercial, standardised and not always user-friendly services – and how it is in demand everywhere. On the occasion of the anniversary, we interview Prof Helmut Reiser, member of the DFN board, deputy director of the LRZ and outspoken fan of well-designed and well-functioning networks.
The DFN association is celebrating its 40th anniversary and you have recently been appointed to its board. Why did you become involved?
Prof. Helmut Reiser: Because the DFN association is exciting, and so is my task. Together with the 354 members of the association and my colleagues on the board of directors and the executive committee, I can actively promote the development of a scientific network that is one of the most powerful and largest in the world and through which more than 1.3 million terabytes or 1.3 exabytes of data are exchanged every year. Networks are my passion - ever since I was a student I have been involved in building them and providing related services such as virtual private networks, user authentication and identification, or security services. Working for the DFN also helps the LRZ to make progress in terms of content, as the challenges facing both institutions are very similar. Just like the DFN, we address the different needs of science and research. We also operate a research network in the form of the Munich Scientific Network, albeit on a smaller scale. We have to be as flexible as possible when dealing with projects where data rates can explode, or when protecting the network and traffic from spam floods known as Distributed Denial of Service or DDoS and other attacks. These are all exciting tasks that I enjoy working on. In addition, there are the international collaborations and projects of the DFN association, which give me insights into areas that would have remained closed to me.
How did you become a part of the DFN?
Reiser: When I joined the LRZ in 2005, it was already clear that I would eventually become involved with the DFN. In 2012, I followed my predecessor into the organisation's Operations Committee. This committee prepares and discusses technical issues related to the DFN. A few years later, as head of the LRZ’s department for communication networks known as KOM, I started to represent the LRZ as a member, and at the end of 2023 I joined the executive board. The relationship between the LRZ and the DFN has always been very close and has a long history.
Tell us why.
Reiser: Technical and, above all, personal connections have played a major role. Communication networks are part of the DNA of the LRZ, which provides services to several universities and colleges that have departments and facilities scattered around the city. The first data lines were installed here in the late 1970s to connect terminals at different locations to the LRZ’s computer. At that time, the architectures were very proprietary, with each manufacturer contributing their own. At the insistence of the then director of the LRZ, Prof Heinz-Gerd Hegering, and with the help of the LRZ, standardised network protocols were introduced. The first network investment programme for the expansion of the network at the universities in Munich was launched in 1986. The LRZ had a component to which all universities were connected. This was the birth of the Münchner Wissenschaftsnetz or Munich Scientific Network, for short MWN. It was also at this time, in 1984, that the DFN association was founded.The LRZ became a member a year later, Hegering was involved in the DFN operating committee, became a member of the DFN board in 1996 and drove the development of a national research and science network from this side. Later, the Bavarian university network and the Munich Scientific Network were integrated into the network.At that time, the technology of all these networks was essentially the same; behind them was the DatexP service of the then monopolist Telekom, which initially slowed down the self-administration process of the first German science network, the DFN’s WiN. Today, universities and research institutions in Munich and Bavaria, as well as national, European and international institutions, can be reached and connected to the Internet via MWN and X-WiN. Almost all services of the LRZ depend on X-WiN. The DFN andles the data exchange with other large network operators such as Telekom, Google and Microsoft, a process known as peering. The LRZ and the MWN are connected to the X-WiN via several optical fibres, and security would not work at all without the DFN: for example, if someone tries to attack our network using a denial of service, the DFN will reject this malicious traffic in its backbone following onsultation with the LRZ. If it were to reach our systems and the necessary security components were located at the LRZ, our lines to X-WiN would be blocked and we would be unable to do anything.
- Founded: 1984
- Members: 354, this are 253 universities
92 research institutes, 9 enterprises - Net: 10.250 km optic fibre connetc 65
Cor net locations und 849 universities
or research institutes - 261 peerings as bridges to the internet,
13 direct peerings with net operators - 1.314.858,88 Terabyte datas were shared
2023 via the X-WiN
This is an exciting story. The DFN has built up X-WiN in a very clever way - 10,250 kilometres of optical fibre and 65 sites are arranged in several rings around a network of eight core nodes, one of which is in Garching. At the LRZ? What does the LRZ contribute?
Reiser: No, the data centre of the Max Planck Society in Garching is where the core network of the DFN is located. The LRZ is directly connected to it via several optical fibres. Core network locations offer the DFN a room with the appropriate infrastructure, such as an uninterruptible power supply, in which there are usually three cabinets: The lines from the commercial providers arrive in one, from there they go via fibre to the DFN equipment, and the third cabinet contains the security systems. Only DFN staff have access to these cabinets, but the site operators enjoy a direct connection to the DFN backbone and therefore very fast data lines.
What are the advantages of this setup?
Reiser: Speed and flexibility. Things are moving fast at the DFN - if we need a 100 Gigabit line to the Regional Computing Centre at the Friedrich-Alexander University in Erlangen, it can be set up quickly. The DFN has also supported us with special cuts, such as a fibre-optic line for a redundant LRZ connection to the university in Freising-Weihenstephan. The DFN backbone is also designed to provide reserves in case data rates explode during a major research project, for example. For example, our high-performance platform for analysing satellite data terrabyte generated massive amounts of traffic when Copernicus data was copied from Poland. Thanks to the monitoring, DFN staff noticed this immediately and increased the peerings on their own initiative. It would be difficult for a commercial network operator to provide this kind of flexibility and support. Companies aim for a high degree of standardisation in their services. Similar to your mobile phone provider, they offer services or service packages for fixed terms, while special requests are very expensive.
How is the DFN finance itself?
Reiser: Through fees for the use of services, through research projects and through contributions from its 354 members – from large universities of excellence to specialised small universities, from research institutes to computer centres and research-related companies. These funds are used to finance the network and the services, their operation and ongoing development, as well as the organisation itself. About 120 IT specialists work for the DFN and DFN-CERT.
DFN-Verein and LRZ are service providers for science and research, do the requirements differ significantly from those of companies?
Reiser: Yes, the requirements of researchers are highly specialised, they are very different and the projects also change quickly. Flexibility and speed are also decisive criteria for X-WiN and DFN-Verein in view of the heterogeneous member community.
We are currently reaching the limits of computer technology, are network capacities also finite?
Reiser: Signal coding techniques and the use of the available transmission spectrum for fibre optics have become better and better. There are limits, of course; I can't put more signals through a line than its spectrum allows, but by cleverly mapping bits or digital information onto analogue signals such as light or electricity, it has been possible to increase bandwidths enormously. At the moment, it doesn't look like we're going to hit a limit any time soon; the DFN association is currently upgrading its backbone to transmission rates per fibre or channel of 400 to 800 gigabits per second.
There are 354 members with very different needs. How do you keep them all happy?
Reiser: There are sometimes heated discussions about technical needs during the general meetings (laughs). But there are several committees in which member institutions, regardless of their size, participate and in which we try to discuss critical issues, get to know the arguments and get a picture of what the members need and want. The Strategic Advisory Board, which also includes my colleague Dieter Kranzlmüller, head of the LRZ, deals with the strategic direction of the DFN; the Operations Committee deals with the technology and operation of the network and services; and the ARuS, the Legal and Security Committee, deals with legal issues, data protection and network security. Based on the opinions and feedback from the committees, the agenda and topics for discussion at the meetings can be prepared. They often focus on how universities can benefit from technology or specific services, but of course also on fees and the fee structure. Each member pays for its DFN connection depending on the bandwidth, but there is cost equality across the board: a connection with the same bandwidth always costs the same in Cologne, Munich, on the Zugspitze or in small towns, even though the Zugspitze and some regions require more effort.
The administration Board of DFN (from left to right): Dr. Hartmut Plehn, Dr. Frank Jenko, Dr. Helmut Reiser,
Kerstin Bein, Christian Zens, Dr. Holger Marten, Dr.-Ing. Stefan Wesner, Dieter Lehmann, Dr. Lars Köller,
Dr. Wolfgang zu Castell, Ilona Glaser, Peter Gietz, Dr.-Ing. Günter Schäfer (o. Abb.) Foto: Frank Homann
How do you work together on the DFN board?
Reiser: Each board member is assigned to a committee, Stefan Wesner heads the Strategy Committee, Christian Zens the ARuS and I the Operations Committee. We make decisions together and coordinate closely so that everyone is involved in the work. We meet every three months in person or online, plus two Operations Committee meetings a year and the two general meetings in Berlin in the summer and Bonn in the winter. There are also special events such as the University Chancellors' Discussion Forum or meetings with ministries. The Secretariat prepares all meetings, discussion papers and draft resolutions very thoroughly. This is very efficient and very valuable for all of us.
Useful net: X-WiN and its 65 core node locations connect 849 universities,
research institiutes and enterprises. Grafik: DFN-Verein
The network is now up and running, it connects 849 locations, and it transports more and more data ever faster. But what will be the challenges for the DFN association in the future?
Reiser: There will certainly be quite a few. With the help of DFN's European counterpart, GÉANT, the networking of research communities in Europe and thus the internationalisation of national science networks is progressing. For example, we are working together in research projects on common protection and security measures in Europe and worldwide, as well as on issues of data protection and data security, and on standards for user authentication and identification. The idea is that if you, as a user of the LRZ in Garching, need certain IT services from a university in Germany, in Europe or perhaps somewhere else in the world, you should be able to access them with your LRZ ID – this is not an easy task, from either a technical or a legal point of view.The next step, which is currently being explored, is a so-called edu ID, a unique identifier or identification that accompanies you throughout your life, even through different university and research institutions. While the DFN was primarily concerned with technology and route optimisation in the first decades after its foundation, security issues became more of a focus at around the turn of the millennium. Initially, the DFN association just issued warnings, but by 2003 it had set up its own Computer Emergency Response Team, DFN-CERT, which protects the network, services and data transfer with increasingly complex IT security services. The core of our work is to improve security for the DFN and its participants, and to provide services for science and, of course, for X-WiN, its continued development and financing. The optical fibre infrastructure is the fundamental basis for everything. The technical components of the network should be modernised every five or six years, at a cost of millions. As things stand at the moment – and of course the DFN has a strategic financial plan – we have sufficient funds to operate, modernise and expand X-WiN and its services into the next decade. This is the only way we can offer our members a cost-neutral increase in bandwidth.
How would you like to go down in the history of the DFN?
Reiser (laughing): I don't think it would be a bad idea if the geo-redundant connection of the core network locations to the DFN fibre-optic network, which we are currently preparing on the board, were to be linked to my work one day. At the moment, the fibre-optic connection to the DFN arrives at a single point in a core network location. However, it would also be more secure in terms of availability and resilience of services if the operators had another physically separate connection in addition to the main one. If one core network node were to fail, the other remote node could take over immediately, and the core network sites could provide the network and services more reliably with the association. We are currently experiencing the need for these redundancy strategies: Bavaria and southern Germany are experiencing massive flooding. A flood or fire in a core network node could have catastrophic consequences. As a fan of good network technology with a soft spot for resilience and security, I could really sleep better if a connection to the DFN fibre-optic network were established at two locations around a core network location…. (Interview: S. Vieser/S. Schlechtweg/LRZ)
Prof Helmut Reiser, board member at DFN and deputy director
at Leibniz Supercomputing Centre.