Communication and Interaction Platform

• Applicant: Professor Dr. Lars Christian Wolf
• Subject Area: Computer Architecture, Embedded and Massively Parallel Systems ; Security and Dependability, Operating-, Communication- and Distributed Systems
• Term: from 2013 to 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 206480214

In phase 1 of CCC, our main contributions in this project have been twofold: First, we built a generic networking platform based on the Stream Control Transmission Protocol (SCTP). The platform provides a wide range of configurations including usage of multiple links for failover or Concurrent Multipath Transfer (CMT). For CMT, we performed a qualitative evaluation that will be complemented by a quantitative analysis especially for mobile scenarios. After completion, CMT with the best mobile performance will be implemented at the end of phase 1.Secondly, we collaborated with further subprojects on contracts in CCC: We developed a comprehensive specification of attributes needed by all considered non-functional viewpoints, and defined the contract negotiation process as well as the architecture implications regarding the Multi-Change Control Layer (MCCL).In phase 2, CCC targets open distributed settings with cooperating systems that communicate using open networks. Especially safety-critical components (e. g. in the MCCL) operating distributedly to achieve common goals will demand tighter requirements for their communication, e. g. for delays or redundancy. We will enhance our platform to support such distributed protocols and their needs. Use cases will be the agreement protocols used by a subproject to achieve Byzantine Fault Tolerance (BFT) and components of distributed Multi-Change Controllers (MCCs) by a subproject. Additionally, a dynamic group management will be provided to support discovery and maintenance of groups of systems.Further, we will focus on increasing the reliability of our platform by employing (1) dense, multi-layered context information regarding network links and (2) contracts for communication.For (1), we will apply the connectivity map concept to all available network links of CCC systems. By geographically collecting their characteristics, upcoming communication properties given the trajectory of a system should be predictable. The gathered information will also be provided to the application subprojects, e. g. contributing to the self-representation of the abilities of vehicles. We will implement advanced strategies to optimize the usage of network links towards different, possibly orthogonal objectives such as costs (energy, monetary) or reliability and redundancy in collaboration with a subproject.For (2), we will extend the work on phase 1 contracts which are relatively static towards applying contracts to the highly dynamic network domain by using multi-level communication contracts for components jointly with a subproject which will provide the theoretical methods.Finally, we will setup two evaluation facilities jointly with a subproject using the domain of an other subproject: (1) A small-scale hardware testbed comprised of vehicular nodes for real-world deployments. (2) A renowned vehicular simulation framework for the evaluation of large scale deployments.

• DFG Programme: Research Units
• Subproject of: FOR 1800: Controlling Concurrent Change (CCC)