SCALABLE DATA PROCESSING IN AUV ONBOARD DISTRIBUTED COMPUTING SYSTEM

Development of distributed computing systems (DCS) takes an important place in modern scientificand technical literature. Generally, only one the most significant features DCS is discussedin papers, for example, performance, reliability, fault tolerance, energy efficiency andscalability. In this paper authors attempt to overall consider problem design, based on theexample a multi-channel onboard data processing places autonomous underwatervehicle (AUV). The aim formulate unified concept onboardDCS real-time processing. As result, architecture and principles operation are proposed, well-known approaches tolerance efficiency, taking into account scalable systems. proposed solutions can beviewed as advancement traditional development. Fault toleranceis achieved by using test-based diagnostic tools. order reduce complexity these tools,redundancy preliminarily added each software module (SM) system. Then tests theredundancies built. It shown that test detects failures addressing exchanges betweenSM blocks implement Based results analysis diagnostictool reaction test, failed detected. stops itswork, new SM implements same algorithm started execute instead. Energy efficiencyproposals suitable case presence redundant processors system which couldsupport multi-core technology. These could be involved execution process thesystem SMs with simultaneous decrease clock frequency supply voltage. Since powerconsumption significantly depends voltage, it decreases alongwith parameter values. An optimal greedy used solve described problem, whichassumes sequential involving additional energyefficiency proposal provides latter capabilities. practicalimportance consists possibility not AUVs application.It also other cases development withreal-time have requirements.