In this report we study the problem of scheduling messages between two parallel machines connected by a low latency network (LAN for instance). The problem of scheduling messages appears in code coupling applications when each coupled code has (at a given state of the simulation) to redistribute the data through a network that cannot handle all the communications at the same time (the network is a bottleneck). We compare two approaches. In the first approach no scheduling is performed. Since all the messages cannot be transmitted at the same time, the transport layer has to manage the congestion (we call this approach the brute-force approach). In the second approach we use two higher-level scheduling algorithms proposed in our previous work called GGP and OGGP. We propose a modelization of the behavior of both approaches and show that we are able to accurately predict the redistribution time with or without scheduling. Although, when the latency is low, the transport layer is very reactive and therefore able to manage the contention very well, we show that the redistribution time with scheduling is always better than the brute-force approach (up to 30%).

EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Hal/Checkpoint

HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 004D84 | SxmlIndent | more

HfdSelect -h $EXPLOR_AREA/Data/Hal/Checkpoint/biblio.hfd -nk 004D84 | SxmlIndent | more

{{Explor lien
   |wiki=    Wicri/Lorraine
   |area=    InforLorV4
   |flux=    Hal
   |étape=   Checkpoint
   |type=    RBID
   |clé=     Hal:inria-00070642
   |texte=   Modelizing, Predicting and Optimizing Redistribution between Clusters on Low Latency Networks
}}