Home > Research > Priority Projects in Research > Collaborative Research Centre 901

Collaborative Research Centre 901

Individualised IT Services in Dynamic Markets

Today, we find ourselves at the start of a new era in the devel­opment and implementation of IT services. We are witnessing the beginnings of a shift away from the 40-year-old principle of either acquiring software by purchasing expensive, relative­ly inflexible standard solutions or relying on the even more expensive method of commissioning customised solutions from external software companies or in-house software departments. With Grid and Cloud Computing, it is now possible to purchase IT services and their essential resources only when necessary and only in the required form. The service-oriented architectures provide methods to put together software at the in-house level, at a minimum. These initial advances towards a new way of providing IT services are the starting point for the research activities in the CRC 901 “On-The-Fly Computing”.

The vision of “On-The-Fly Computing” is one of individually and automatically configured and implemented IT services, consisting of flexibly combinable services that are available on free markets. At the same time, CRC 901 is aimed at organising markets whose participants maintain a lively service landscape by dedicated entrepreneurial action. With this vision, CRC 901 looks far into the future of IT development and usage, the transformation of which we are already experiencing today.
In order to research the extent to which this vision can be realised, CRC 901 will develop methods and techniques that

  • enable an almost entirely automatic configuration, implementation and adaptation of IT services from the services available on worldwide markets,
  • guarantee the protection not only of the services acquired in this way, but also of the active participants in themarkets, and
  • support the organisation and further development of these markets and the necessary interaction between those involved.

To reach these goals, computer science experts from diverse disciplines, such as software technology, algorithmics, computer networks, system design, security and cryptology are working hand-in-glove with economists who contribute their specific expertise on how to promote the organisation and the further development of the market.

On an organisational level, CRC 901 is represented by its Executive Board, consisting of Professor Friedhelm Meyer auf der Heide as Chairman with Professor Heike Wehrheim, Professor Marco Platzner and Professor Claus Jochen Haake as Deputy Chairmen and Dr. Ulf-Peter Schroeder as Executive Officer. In total, four workgroups from the Heinz Nixdorf Institute, eight university chairs from Computer Sciences, five university chairs from the Faculty of Economics, and four junior professors are involved in the CRC.

First Funding Period for CRC 901: July 2011 – June 2015

During the first funding period, we made significant progress towards demonstrating the feasibility of On-The-Fly Computing, developing the methods and techniques required to solve critical sub-issues. We developed description languages for services that contain sufficient information to allow service configurations as well as initial methods for an automatic configuration process. To support the resource-efficient execution of configured services, we designed scheduling algorithms in heterogeneous execution environments. We also developed verification and test methods that enable the quality of the configured IT services to be assessed. To organise the strategic behaviour of the actors in On-The-Fly Computing (i.e. the users who request services, the providers who configure these services from sub-services and the suppliers of sub-services and execution resources), we conducted economic studies on the costs and reputation development. To support the interaction between the actors necessary for the above tasks, we developed peer-to-peer based methods that support the interaction needs of the actors by adapting the overlay network underneath the peer-to-peer system.

Various publications about single research issues and prototypical developments of tools and demonstrators convinced the DFG reviewers from the state of the CRC. Professor Meyer auf der Heide explains: “With the renewal of funding by the DFG, we can take the next step and, inter alia, develop concepts for the user-friendliness, safety and cost-effectiveness of On-The-Fly Computing. Thus, the vision becomes a competitive paradigm for future markets of IT services.”

Second Funding Period for CRC 901: July 2015 – June 2019

In the second funding period, we will focus on some new basic research issues. This includes further developing the user friendliness of the description languages, achieving greater involvement of machine learning methods in the configuration and quality assessment of service configurations, increasing the efficiency of OTF Compute Centres through re-configurability, researching the dynamics of the actors’ interaction requirements in the markets and overlay networks, and increasing the use of empirical methods to analyse and influence the strategic behaviour of the actors. We will continue to intensify our project-wide discussions, research and development of architectures, demonstration systems and prototypes. To organise these activities, we will set up cross-cutting issues orthogonally to the research in the individual subprojects. One main focus during the current funding period will be to create closer ties between the economic sciences and computer science.

Another main focus for the second funding period is the exemplary implementation of a variant of an on-the-fly software architecture within a testbed. This testbed is primarily intended to serve as an experimental and evaluation platform for the CRC subprojects with respect to their implemented components. In addition, this testbed may also be used as integral demonstrator concerning the feasibility of on-the-fly computing for the evaluation at the end of the second phase.

On 10 November 2017 we discussed the development status of the TestBed and an application scenario of OTF Computing in the field of machine learning with the advisory board of the CRC. With this on-the-fly machine learning (OTF-ML) concept, we want to make a veritable contribution to the developments that have become established in the scientific community under the term AutoML - Machine Learning for Automated Algorithm Design. At the advisory board meeting, we discussed also about questions concerning the stabilisation and further industry transfer options of the on-the-fly computing paradigm.