"Extending the Reach of Middleware"

Extending the Reach of Middleware" is our mission statement.

Within this context, the Operating Systems and Middleware Group at HPI (Prof. Dr. rer.nat. habil. Andreas Polze) focuses on wide-area distributed computing (Grid Computing) following the service-based system engineering approach on one hand side, as well as on interconnecting middleware-based systems and embedded computing environments. Our research aims at new algorithms, paradigms, design patterns, and implementation techniques for predictable service-based systems.

For the HPI research college, we propose the following research questions:

• Dynamic Reconfiguration in Adaptive Systems
  Research in dynamic reconfiguration of sevice-oriented and component-based software focuses on design, implementation, and experimental evaluation of algorithms and techniques for application adaptation at runtime to deal with changing environments. The main challenge with runtime reconfiguration is guaranteeing certain functional properties of the application under configuration. System consistency and existing communnications relations have to be taken into account.
• Embedded Systems in the Distributed Control Lab
  The Distributed Control Lab investigated the interconnection of middleware-based components and embedded systems. Our research deals with paradigms and algorithms for pedictable system behavior in instable environments at the "edge of the grid". We focus on non-functional system properties, such as real-time, fault-tolerance, safety, and security. Experiment hardware has to be protected against potentially malicious control algorithms using combined software and hardware mechanisms. Structuring of embedded systems in a service-oriented world is an open research question.
• Predictable Grid-Computing - Adaptive Services Grid - European Project
  The Adaptive Services Grid (ASG) - Project is an integrated project in the 6th framework of the European IST program in the area "open development platforms for software and services". ASG has started on September 1, 2004 and is comprised of 21 partners from 7 nations.
  The Operating Systems and Middleware Group at HPI leads the workk component "Services Grid Infrastructure" (C-5). C-5 designs and implements the Adaptive Execution Platform (AXP) for the ASG-project and investigates techniques for ensuring predictable system behavior in terms of certain, non-functional system properties (timing behavior, availability, reliability, security) in heterogeneous systems. The definitions and implementation of description techniques for meta-data in service-based systems is an open research question.
• Distributed Coordination - Das Grid-Occam Projekt
  Occam is a parallel processing language designed by a team at INMOS in conjunction with the design of the transputer processor, and based on Sir T. Hoare's ideas of Communicating Sequential Processes (CSP). Occam incorporates support for very fine grained, easy to use threads and seamless support of multi-processor environments. It can be used with shared or distributed memory systems, and the strong basis in CSP makes it excellent choice when formal proofs of correctness are required - the application of Occam as a configuration language for wide-area distributed computing is an open research question.
• Aspect-oriented Programming - The LOOM.NET Project
  Aspect-oriented programming is a relatively new, promising approach to modelling and implementation of distributed systems with predictable system behavior. Our research investigates the usage of aspects as a mechanism to express required non-functional component and system properties. Tools, such as an aspect weaver, may use aspect information to extend component code in order to guarantee certain system properties. In context of the HPI research college, we propose the investigation of new approaches to aspect-oriented programming and its implementation and evaluation in context of predictable, service-based systems.