Performance Evaluation of a Distributed and Probabilistic Network Monitoring Approach

QC 20140509
We investigate the effects of employing a proba- bilistic fault detection approach relative the performance of a deterministic network monitoring method. The approach has its foundation in probabilistic network management, in which performance limits and thresholds are specified in terms of e.g. probabilities or belief values. When combined with adap- tive mechanisms, probabilistic approaches can potentially offer improved controllability, adaptivity and reliability, compared to deterministic monitoring methods. Results from synthetically generated and real network QoS measurements indicate that the probabilistic approach generally can perform at least as good as a deterministic algorithm, with a higher degree of predictable performance and resource-efficiency. Due to the stochastic nature of the algorithm, worse performance than expected is sometimes observed. Nevertheless, the results give additional support to some of the practical benefits expected in using probabilistic approaches for network management purposes.

Gillblad, DanielRISE,Decisions, Networks and Analytics lab,SICS(Swepub:ri)daniel.gillblad@ri.se (författare)
RISEDecisions, Networks and Analytics lab (creator_code:org_t)

Om ämnet

TEKNIK OCH TEKNOLOGIER: TEKNIK OCH TEKNO ...; och Elektroteknik oc ...; och Kommunikationssy ...

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.