Avoiding Network Congestion with Local Information

IEEE Trans. Parallel Distrib. Syst.

Duato

2005

Self Cite

Multiprocessor interconnection networks may reach congestion with high traffic loads, which prevents reaching the wished performance. Unfortunately, many of the mechanisms proposed in the literature for congestion control either suffer from a lack of robustness, being unable to work properly with different traffic patterns or message lengths, or detect congestion relying on global information that wastes some network bandwidth. This paper presents a family of mechanisms to avoid network congestion in wormhole networks. All of them need only local information, applying message throttling when it is required. The proposed mechanisms use different strategies to detect network congestion and also apply different corrective actions. The mechanisms are evaluated and compared for several network loads and topologies, noticeably improving network performance with high loads but without penalizing network behavior for low and medium traffic rates, where no congestion control is required.

Section: Discussionmentioning

confidence: 99%

“…On the other hand, each node has four injection/ejection channels. We showed the advantages of having several injection/ejection channels in [3].…”

Section: Network Modelmentioning

confidence: 95%

A family of mechanisms for congestion control in wormhole networks

Baydal

IEEE Trans. Parallel Distrib. Syst.

Duato

2005

Self Cite

“…Moreover, it progressively reduces the bandwidth available to inject messages, either by reducing the number of injection channels or, if it is necessary, by applying increasing intervals of forbidden message injection. The mechanism has been exhaustively tested, achieving good results in all the evaluated conditions and improving recent proposals [2], [15].…”

Section: Introductionmentioning

confidence: 96%

“…Message throttling [4] (stop or slow down the injection rate of messages) has been the most frequently used method. Several mechanisms for multiprocessors have already been proposed but they have important drawbacks [2]. In this paper, we propose and evaluate a new mechanism to avoid network saturation that tries to overcome these drawbacks.…”

Section: Introductionmentioning

confidence: 99%

A Robust Mechanism for Congestion Control: INC

Baydal

Euro-Par 2003 Parallel Processing

2003

Several techniques to prevent congestion in multiprocessor interconnection networks have been recently proposed. Unfortunately, they either suffer from a lack of robustness or detect congestion relying on global information that wastes a lot of transmission resources. This paper presents a new mechanism that uses only local information to avoid network saturation in wormhole networks. It is robust and works properly in different conditions. It first applies preventive measures of different intensity depending on the estimated traffic level; and if necessary, it uses message throttling during predefined time intervals that are extended if congestion is repeatedly detected. Evaluation results for different network loads and topologies show that the proposed mechanism avoids network performance degradation. Most important, without introducing any penalty for low and medium network loads.

“…In [16], it was shown that deadlocks rarely occur when sufficient routing freedom is provided, but are more likely to occur when the network is close to or beyond saturation. Moreover, the use of techniques that avoid network saturation such as the message injection limitation mechanisms proposed in [14], [4], [5] reduces the probability of deadlock to negligible levels, even when fully adaptive routing is used with only a few virtual channels. These results suggest that deadlock recovery techniques can be more effective than avoidance techniques because it is more worthwhile to use the available resources to supply unrestricted fully adaptive routing than to avoid deadlocks that seldom occur.…”

Section: Introductionmentioning

confidence: 99%

Fc3d: flow control-based distributed deadlock detection mechanism for true fully adaptive routing in wormhole networks

Rubio

IEEE Trans. Parallel Distrib. Syst.

Duato

2003

Self Cite

Two general approaches have been proposed for deadlock handling in wormhole networks. Traditionally, deadlockavoidance strategies have been used. In this case, either routing is restricted so that there are no cyclic dependencies between channels or cyclic dependencies between channels are allowed provided that there are some escape paths to avoid deadlock. More recently, deadlock recovery strategies have begun to gain acceptance. These strategies allow the use of unrestricted fully adaptive routing, usually outperforming deadlock avoidance techniques. However, they require a deadlock detection mechanism and a deadlock recovery mechanism that is able to recover from deadlocks faster than they occur. In particular, progressive deadlock recovery techniques are very attractive because they allocate a few dedicated resources to quickly deliver deadlocked messages, instead of killing them. Unfortunately, distributed deadlock detection is usually based on crude time-outs, which detect many false deadlocks. As a consequence, messages detected as deadlocked may saturate the bandwidth offered by recovery resources, thus degrading performance. Additionally, the threshold required by the detection mechanism (the time-out) strongly depends on network load, which is not known in advance at the design stage. This limits the applicability of deadlock recovery on actual networks. In this paper, we propose a novel distributed deadlock detection mechanism that uses only local information, detects all the deadlocks, considerably reduces the probability of false deadlock detection over previously proposed techniques, and is not significantly affected by variations in message length and/or message destination distribution.