Abstract:Abstract:Computing systems are becoming more and more dynamically reconfigurable or adaptive, to be flexible w.r.t. their environment and to automate their administration. Autonomic computing proposes a general structure of feedback loop to take this into account. In this report, we are particularly interested in approaches where this feedback loop is considered as a case of control loop where techniques stemming from Control Theory can be used to design efficient safe, and predictable controllers. This approa… Show more
“…Control theory has been applied in High-performance Computing rather scarcely therefore it is still in need for basic research [3], [4], [5]. Most previous approaches propose rather simple linear models that, to our best understanding, lack the ability to reflect several nonlinear behaviors and variable constraints of the real system, since their accuracy is limited to a specific region of operation, failing to ensure the desired performance and stability objectives for all possible situations.…”
Section: A Control Theory For High Performance Computingmentioning
“…Control theory has been applied in High-performance Computing rather scarcely therefore it is still in need for basic research [3], [4], [5]. Most previous approaches propose rather simple linear models that, to our best understanding, lack the ability to reflect several nonlinear behaviors and variable constraints of the real system, since their accuracy is limited to a specific region of operation, failing to ensure the desired performance and stability objectives for all possible situations.…”
Section: A Control Theory For High Performance Computingmentioning
“…Recently, the perspectives of self-adaptive systems and autonomic management have also been considered in HPC. AC relies on a structure of feedback loop, where a controller, separate from the system to be managed, runs in parallel to it: the well known MAPE-K loop [4] involves components to Monitor (through probes and sensors), Analyze (extracting relevant information from data, and take decisions), Plan (transform decisions into adaptation actions) and Execute (perform adaptation actions). All this is done using Knowledge on the system (reified representation, data-base).…”
Section: Background a Autonomic Computing And Its Controlmentioning
HPC systems are facing more and more variability in their behavior, related to e.g., performance and power consumption, and the fact that they are less predictable requires more runtime management. This can be done in an Autonomic Management feedback loop, in response to monitored information in the systems, by analysis of this data and utilization of the results in order to activate appropriate system-level or application-level feedback mechanisms (e.g., informing schedulers, down-clocking CPUs). One such problem is found in the context of CiGri, a simple, lightweight, scalable and fault tolerant grid system which exploits the unused resources of a set of computing clusters. Computing power left over by the execution of a main HPC application scheduling is used to execute smaller jobs, which are injected as much as the global system allows. This paper presents first results addressing the problem of automated resource management in an HPC infrastructure, using techniques from Control Theory to design a controller that maximizes cluster utilization while avoiding overload. We put in place a mechanism for feedback (Proportional Integral, PI) control system software, through a maximum number of jobs to be sent to the cluster, in response to system information about the current number of jobs processed.
“…Other perspectives are in co-ordination of multiple autonomic loops, switching controllers (integrating into the DSL schemes explored in another context [7]) ; modularity and hierarchical loops, both for re-use and to managed compilation complexity, and considering more ad-vanced discrete control features like logico-numeric properties [8] ; and integrating the DSL for Autonomic Managers into a global design process for FPGA-based applications. We also consider the general perspectives of basing Autonomic Management of Autonomic Computing systems on Control Theory [9,10].…”
in the field of Autonomic Computing. He currently works on the modelbased control of adaptive and reconfigurable computing systems, using techniques from Control Theory, and ranging from embedded systems, to Cloud distributed systems and High-Performance Computing.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.