TY - JOUR
T1 - Impact of tree structure on retransmission efficiency for TRACK
AU - Busson, Anthony
AU - Rougier, Jean Louis
AU - Kofman, Daniel
PY - 2001/1/1
Y1 - 2001/1/1
N2 - This paper focuses on tree based reliable multicast protocols, or more precisely TRACK (Tree based Acknowledgment) protocols, as defined by IETF. With the TRACK approach, the classical feedback implosion problem is handled by a set of servers, organized in a tree structure, which are in charge of local retransmissions and feedback aggregation. We study the impact of the control tree structure (for instance the number of servers to be deployed) on transmission performances. We propose a new model, where point processes represent the receivers and the servers, which captures loss correlation phenomena. We are able to get explicit expressions of the number of useless retransmissions (useless as the given segment was already received) as a function of a limited number of tree characteristics. Generic tree configuration rules, optimizing transmission efficiency, are obtained.
AB - This paper focuses on tree based reliable multicast protocols, or more precisely TRACK (Tree based Acknowledgment) protocols, as defined by IETF. With the TRACK approach, the classical feedback implosion problem is handled by a set of servers, organized in a tree structure, which are in charge of local retransmissions and feedback aggregation. We study the impact of the control tree structure (for instance the number of servers to be deployed) on transmission performances. We propose a new model, where point processes represent the receivers and the servers, which captures loss correlation phenomena. We are able to get explicit expressions of the number of useless retransmissions (useless as the given segment was already received) as a function of a limited number of tree characteristics. Generic tree configuration rules, optimizing transmission efficiency, are obtained.
U2 - 10.1007/3-540-45546-9_9
DO - 10.1007/3-540-45546-9_9
M3 - Article
AN - SCOPUS:84867439002
SN - 0302-9743
VL - 2233
SP - 113
EP - 127
JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ER -