Paper Info
Title
Packet Dropping and Decoding Statistics for Linear Network Coded Broadcast with Feedback
Abstract
Exploitation of the benefit of feedback in network coded multicast or broadcast is a well-investigated technique, whose purpose is to make the system more robust to loss. In such scenario, the drop when seen algorithm reduces the expected queue length at sender while maintaining throughput optimality of encoding scheme. In this work, we study a broadcasting system where the transmitter is sending data packets of a generation, using drop when seen strategy, to a set of receivers of finite buffer length over packet erasure channels with full feedback. We develop an n-dimensional Markov chain-based model that enables us to compute the probability of dropping a packet (which is currently being transmitted) from sender queue at the end of a particular time slot. Further, by proposing the idea of first time dropping, we map the n-dimensional dynamics to a simple one-dimensional model. This simplified framework is used to find expected time to drop a packet from the sender queue. It also provides an opportunity to study decoding statistics of stored packets in receivers' buffer. We analyze three different decoding options namely, arbitrary single receiver decoding, leader decoding, worst follower decoding, and compare them.
Year
DOI
Venue
2018
10.1109/LCNW.2018.8628590
2018 IEEE 43rd Conference on Local Computer Networks Workshops (LCN Workshops)
Keywords
Field
DocType
Broadcast,Decoding probability,Drop when seen,Dropping probability,Network coding,Sender queue
Linear network coding,Broadcasting,Transmitter,Computer science,Queue,Network packet,Computer network,Throughput,Decoding methods,Multicast,Statistics
Conference
ISBN
Citations 
PageRank 
978-1-5386-5098-1
0
0.34
References 
Authors
9
2
Name
Order
Citations
PageRank
Sovanjyoti Giri100.68
Rajarshi Roy28018.39