Institut für Parallele und Verteilte Systeme (IPVS)

Publikationen

Eine Übersicht der Publikationen des Instituts für Parallele und Verteilte Systeme

Publikationen VS: Bibliographie 2021 BibTeX

 
@inproceedings {INPROC-2021-01,
   author = {David Hellmanns and Lucas Haug and Moritz Hildebrand and Frank D{\"u}rr and Stephan Kehrer and Ren{\'e} Hummen},
   title = {{How to Optimize Joint Routing and Scheduling Models for TSN Using Integer Linear Programming}},
   booktitle = {Proceedings of the 29th International Conference on Real-Time Networks and Systems},
   editor = {ACM},
   address = {Nantes},
   publisher = {ACM (Online)},
   institution = {Universit{\"a}t Stuttgart, Fakult{\"a}t Informatik, Elektrotechnik und Informationstechnik, Germany},
   pages = {1--12},
   type = {Konferenz-Beitrag},
   month = {April},
   year = {2021},
   isbn = {10.1145/3453417.3453421},
   keywords = {Time-Sensitive Networking, TSN, Scheduling, Routing, Integer Linear Programming, Optimization, Model, ILP},
   language = {Englisch},
   cr-category = {D.4.7 Operating Systems Organization and Design},
   ee = {ftp://ftp.informatik.uni-stuttgart.de/pub/library/ncstrl.ustuttgart_fi/INPROC-2021-01/INPROC-2021-01.pdf},
   contact = {david.hellmanns@ipvs.uni-stuttgart.de},
   department = {Universit{\"a}t Stuttgart, Institut f{\"u}r Parallele und Verteilte Systeme, Verteilte Systeme},
   abstract = {Reliable real-time communication is an essential technology for industrial manufacturing but also other branches to transport mission-critical messages. IEEE Time-Sensitive Networking (TSN) is a disruptive real-time communication standard extending IEEE Ethernet with real-time mechanisms. One of the core features of TSN is the Time-Aware Shaper (TAS) enabling TDMA-based scheduling of streams within the network. TDMA has many advantages from the real-time perspective. Foremost, stream isolation in the time dimension enables tight delay and jitter bounds. Moreover, conformance to these bounds is proven by the design of the TDMA schedule. However, calculating an optimal schedule is an NP-hard problem. Therefore, various approaches to optimize the schedule calculation are proposed, such as Integer Linear Programming (ILP). Nevertheless, a systematic comparsion of the different optimization approaches with respect to their performance is missing so far. To fill this gap, we first provide a systematic classification of optimizations of ILP-based TSN scheduling. To quantify the effects of such optimization approaches, we introduce a base ILP and propose optimizations for the different categories. Using the proposed optimization, we evaluate the performance with regard to execution time and schedulability (number of solved schedules). Our results show that the optimizations lead to strongly fluctuating results. Certain intuitive optimizations can even lead to massive performance degradations.},
   url = {http://www2.informatik.uni-stuttgart.de/cgi-bin/NCSTRL/NCSTRL_view.pl?id=INPROC-2021-01&engl=0}
}
@article {ART-2021-02,
   author = {Otto Bibartiu and Frank D{\"u}rr and Kurt Rothermel and Beate Ottenw{\"a}lder and Andreas Grau},
   title = {{Scalable k-out-of-n models for dependability analysis with Bayesian networks}},
   journal = {Reliability Engineering \& System Safety},
   editor = {Paolo Gardoni},
   publisher = {Elsevier Science Ltd.},
   volume = {210},
   pages = {1--13},
   type = {Artikel in Zeitschrift},
   month = {Februar},
   year = {2021},
   isbn = {10.1016/S0951-8320(21)00145-9},
   keywords = {Availability; Scalability; Voting Gate; Fault-Tree; Bayesian networks},
   language = {Englisch},
   cr-category = {B.8.1 Reliability, Testing, and Fault-Tolerance,     C.4 Performance of Systems},
   ee = {ftp://ftp.informatik.uni-stuttgart.de/pub/library/ncstrl.ustuttgart_fi/ART-2021-02/ART-2021-02.pdf,     https://doi.org/10.1016/j.ress.2021.107533},
   contact = {otto.bibartiu@ipvs.uni-stuttgart.de},
   department = {Universit{\"a}t Stuttgart, Institut f{\"u}r Parallele und Verteilte Systeme, Verteilte Systeme},
   abstract = {Availability analysis is indispensable in evaluating the dependability of safety and business-critical systems, for which fault tree analysis (FTA) has proven very useful throughout research and industry. Fault trees (FT) can be analyzed by means of a rich set of mathematical models. One particular model are Bayesian networks (BNs) which have gained considerable popularity recently due to their powerful inference abilities. However, large-scale systems, as found in modern data centers for cloud computing, pose modeling challenges that require scalable availability models. An equivalent BN of a FT has no scalable representation for the k-out-of-n (k/n) voting gate because the conditional probability table that constitutes the k/n voting gate grows exponentially in n. Thus, the memory becomes the limiting factor. We propose a scalable k/n voting gate representation for BNs, based on the temporal noisy adder. The resulting model reduces the initial exponential to polynomial memory growth without a custom inference algorithm. Previous BN implementations of the k/n voting gate could only handle around 30 input events until memory limits make inference infeasible. However, our evaluation shows that our scalable model can handle more than 700 input events per gate, making it possible to evaluate large scale systems.},
   url = {http://www2.informatik.uni-stuttgart.de/cgi-bin/NCSTRL/NCSTRL_view.pl?id=ART-2021-02&engl=0}
}
 
Zum Seitenanfang