Designing and operating computing and communication systems is becoming increasingly challenging due to a multitude of reasons. First, compute nodes evolve towards parallel and heterogeneous architectures to realise performance gains while minimising their power consumption. Progress in micro(nano)-electronics allows us to integrate more and more functionality on a single compute node, but at the same time requires us to deal with increasing numbers of faulty and unreliable components. Second, distributed systems grow in the numbers and heterogeneity of nodes and must be able to cope with an increasing level of dynamics. The network topology and the collective resources of the distributed system can vary strongly during runtime since nodes may be mobile, break, run out of battery, or generally leave and enter the network. Third, future challenging application domains have quite divergent requirements with respect to functionality and flexibility, performance, resource usage and costs, reliability and safety, and security. Fuelled by technological progress, applications with exciting levels of user interaction will be possible. Current design and operation principles and methods will neither be able to scale with future systems, nor handle the variety of requirements and optimisation goals efficiently. Novel design and operation principles and methods are needed, possibly those that radically break with the static ones we have built into our systems and the fixed abstraction layers we have learned to cherish over the years. With the EPiCS project, we propose to study computing and communication systems that collect and maintain information about their state and progress, which enables them to reason about their behaviour (self-awareness) and utilise this knowledge to effectively and autonomously adapt their behaviour to changing conditions (self-expression). We denote the basic ability to collect and maintain information about state and progress as proprioception, referring to psychology where proprioception (from Latin proprius, meaning "one's own", and perception) is defined as the sense of the relative position of neighbouring parts of the body. Proprioception is the enabler for building up awareness and realising advanced autonomous behaviour.
Proprioception, self-awareness, and self-expression are concepts mainly known from psychology, philosophy and medicine. These concepts are rather new to the domains of computing and networking and are thus not yet sufficiently investigated and understood. The innovative claim of EPiCS is that the successful transfer of these concepts to computing and networking domains will help create the powerful and versatile heterogeneous and distributed systems of 2020 and beyond.
In EPiCS we will investigate the transfer of principles from other scientific disciplines to computer science and engineering. Based on the gained insights, we will derive novel design and operation methods and tools. This approach has also been followed for other nature-inspired computing paradigms, such as evolutionary computing and swarm intelligence, which despite their success often lack the assurances and hard guarantees required for the construction of technological systems. EPiCS is therefore devoted to engineering proprioception, self-awareness and self- expression into computing and communication systems, including a thorough analysis of the limits of the approach and its suitability for different application domains. EPiCS intends to reach its ambitious objectives by integrating multidisciplinary research from several areas:
- concepts and foundations for self-aware and self-expressive systems
- hardware/software platform technologies for autonomic compute nodes
- self-aware network architectures and middleware layers
Methodologically, EPiCS follows an iterative approach and drives and validates the foundational and technological research by the requirements of three challenging application domains that cover both high-end computers and embedded systems, as well as embeddings into technical and non-technical contexts: heterogeneous compute clusters for financial modelling, distributed smart cameras for person detection and tracking, and hypermusic on an interactive mobile media system.
This setup will lead to an exciting interaction of theory and practice, and combine long-term foundational research with meeting challenges from promising applications.