We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.PF

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Performance

Title: Towards self-optimization of publish/subscribe IoT systems using continuous performance monitoring

Abstract: Today, more and more embedded devices are being connected through a network, generally Internet, offering users different services. This concept refers to Internet of Things (IoT), bringing information and control capabilities in many fields like medicine, smart homes, home security, etc. Main drawbacks of IoT environment are its dependency on Internet connectivity and need continuous devices power. These dependencies may affect system performances, namely request processing response times. In this context, we propose in this paper a continuous performance monitoring methodology, applied on IoT systems based on Publish/subscribe communication model. Our approach assesses performances using Stochastic Petri net modeling, and self-optimizes whenever poor performances are detected. Our approach relies on a Stochastic Petri nets modelling and analysis to assess performances. We target improving performances, in particular response times, by online modification of influencing factors.
Comments: 13 pages, 7 figures. Conference: 5th International Conference on Networks, Blockchain and Internet of Things (NBIoT 2024). Published by Computer Science Conference Proceedings in Computer Science & Information Technology (CS & IT)
Subjects: Performance (cs.PF)
Journal reference: Computer Science & Information Technology (CS & IT) 2024
DOI: 10.5121/csit.2024.140502
Cite as: arXiv:2404.14926 [cs.PF]
  (or arXiv:2404.14926v2 [cs.PF] for this version)

Submission history

From: Mohammed Djahafi [view email]
[v1] Tue, 23 Apr 2024 11:11:20 GMT (910kb)
[v2] Mon, 29 Apr 2024 11:15:50 GMT (910kb)

Link back to: arXiv, form interface, contact.