Related European projects
Several European projects have been or are concerned with building frameworks, libraries or with the integration of scheduling algorithms, e.g., OCERA, FIRST, OMI/CLEAR. The objectives of PREDATOR are complementary to their goals. PREDATOR is addressing predictability acting on several aspects, including compilers, architecture, kernel, etc.
PDCS2 aimed at an effective development process for dependable distributed real-time systems. A timeliness analysis was included to guarantee real-time constraints.
HIDOORS was concerned with bringing a particular language, Java, to the distributed, real time system domain increasing its predictability properties as to memory management and remote access for distributed real-time programming.
Two projects started from a particular technology, TTA, designed to increase predictability and reliability partly trading performance for predictability:
- nextTTA attempted to further improve TTA’s dependability and functionality. It contained the development of a WCET estimation algorithm, but not the goal to reconcile predictability with performance.
- DECOS aims to increase the dependability of embedded applications mainly based on TTA. A goal related to the ones of PREDATOR concerned the predictability of temporal properties of encapsulated virtual networks. However, there was no attempt to increase predictability while maintaining performance.
CORTEX was concerned with QoS as a metric of predictability in terms of timeliness and reliability.
OCEANS focused on the design of specific compilers for embedded systems. The scope included only very long instruction word (VLIW) machines, mainly the Philips TriMedia processor.
Several European projects are concerned with multi-core architectures:
- ÆTHER aims at self-adaptive computing. This approach, though needed in some embedded systems, adds another level of complexity and endangers predictability.
- SARC aims at a scalable integrated architecture applicable to a wide range of applications. Predictability is not a dominant design goal.
- SHAPES is concerned with scalable architectures obtained through tiling with building blocks.
- HARTES aims at rapid design of embedded systems from high-level descriptions focussing on algorithm exploration tools using reconfigurable hardware technologies.
The German nationally funded project AVACS (Automatic Verification and Analysis of Complex Systems) is among others concerned with validating timing-analysis methods and tools as well as timing properties of embedded systems.
PREDATOR partners are well-connected with the embedded-systems community, on the academic side as well as on the industrial side. These connections will be used to import new insights and techniques from outside of Predator as well as making PREDATOR results known to relevant parties.
PREDATOR has established links with the European Network of Excellence ARTIST2 and the likely successor network ARTIST DESIGN through joint membership of partners. The same holds for AVACS, where Reinhard Wilhelm and Lothar Thiele are proposers for a second funding period.
Airbus, AbsInt, and Saarland University are partners in the ITEA2 project ES_PASS (Embedded Systems — Product-based Assurance), which deals with the introduction of static analysis into industrial processes.
AbsInt is partner in the German nationally funded Verisoft project. This link can be used to exploit PREDATOR results in some industrial applications from the automotive domain. A natural group of customers is the AUTOSAR consortium as they need to deal with resource requirements.
AbsInt also participates in the German nationally funded project SuReal on the creation of a pervasive development process in which safety aspects are continuously taken into account and can be formally verified at any time.
Furthermore, AbsInt participates in ARTIST2, ES_PASS, the European project EmBounded on the analysis and certification of resource needs of programs in a high-level programming language, and the European project INTEREST aiming at overcoming the current lack of integration and interoperability of tools for developing embedded systems software.
University of Bologna is a member of the ARTIST2 and ARTIST DESIGN core consortium and it will thereby further strengthen the strong link between PREDATOR and the research community on embedded systems.
University of Bologna is also a member in the HIPEAC Network of Excellence. This will help establishing links between PREDATOR and the computer architecture community, through presentation in HIPEAC-organised events.
Moreover, University of Bologna coordinates the IST project SensAction-AAL on ambient-assisted living. Thus, it will be possible to establish a link between the work being done in these projects, creating initiatives aimed at establishing the importance of predictable execution in embedded systems for ambient-assisted living.
The Swiss Federal Institute of Technology is involved in the project SHAPES. The purpose of the project is to establish a software and hardware infrastructure for massively parallel embedded systems. As the target applications in signal and image processing require predictability, the methods developed in PREDATOR will be of major use in SHAPES.
Besides its membership in the already mentioned Networks of Excellence ARTIST2 and ARTIST DESIGN, TU Dortmund is a partner in the European Network of Excellence HiPeac and its likely successor network HiPeac2. This connection will be useful for exchanging knowledge on compilation for multi-processor systems.
In addition, TU Dortmund is member of the European STREP project MORE that deals with resource-constrained embedded systems. This link will be valuable since resource management is a central aspect of both PREDATOR and MORE.
TU Dortmund is also cooperating with its local spin-off ICD on the STREP project MNEMEE (Memory management technology for adaptive and efficient design of embedded systems).
Bosch, as member of the AUTOSAR consortium, will maintain links to this consortium.