On unreliable computing systems when heavy-tails appear as a result of the recovery procedure
ACM SIGMETRICS Performance Evaluation Review - Special issue on the workshop on MAthematical performance Modeling And Analysis (MAMA 2005)
ACM SIGMETRICS Performance Evaluation Review
Dynamic packet fragmentation for wireless channels with failures
Proceedings of the 9th ACM international symposium on Mobile ad hoc networking and computing
Is ALOHA causing power law delays?
ITC20'07 Proceedings of the 20th international teletraffic conference on Managing traffic performance in converged networks
File fragmentation over an unreliable channel
INFOCOM'10 Proceedings of the 29th conference on Information communications
Modulated Branching Processes, Origins of Power Laws, and Queueing Duality
Mathematics of Operations Research
Uniform approximation of the distribution for the number of retransmissions of bounded documents
Proceedings of the 12th ACM SIGMETRICS/PERFORMANCE joint international conference on Measurement and Modeling of Computer Systems
Hi-index | 0.00 |
Frequent failures characterize many existing communication networks, e.g. wireless ad-hoc networks, where retransmission- based failure recovery represents a primary approach for successful data delivery. Recent work has shown that retransmissions can cause power law delays and instabilities even if all traffic and network characteristics are super-exponential. While the prior studies have considered an independent channel model, in this paper we extend the analysis to the practically important dependent case. We use modulated processes, e.g. Markov modulated, to capture the channel dependencies. We study the number of retransmissions and delays when the hazard functions of the distributions of data sizes and channel statistics are proportional, conditionally on the channel state. Our results show that the tails of the retransmission and delay distributions are asymptotically insensitive to the channel correlations and are determined by the state that generates the lightest asymptotics. This insight is beneficial both for capacity planning and channel modeling since we do not need to account for the correlation details. However, these results may be overly optimistic when the best state is infrequent, since the effects of 'bad' states may be prevalent for sufficiently long to downgrade the expected performance.