Dynamics of TCP/RED and a scalable control

Abstract

Abstract-We demonstrate that the dynamic behavior of queue and average window is determined predominantly by the stability of TCP/RED, not by AIMD probing nor noise traffic. We develop a general multi-link multi-source model for TCP/RED and derive a local stability condition in the case of a single link with heterogeneous sources. We validate our model with simulations and illustrate the stability region of TCP/RED. These results suggest that TCP/RED becomes unstable when delay increases, or more strikingly, when link capacity increases. The analysis illustrates the difficulty of setting RED parameters to stabilize TCP: they can be tuned to improve stability, but only at the cost of large queues even when they are dynamically adjusted. Finally, we present a simple distributed congestion control algorithm that maintains stability for arbitrary network delay, capacity, load and topology.

Stability analysis of multiple-bottleneck networks

“…; M, the larger the absolute values of kðAÞ and the smaller the kPB i k, and hence the better the chance that the system is asymptotically stable. These observations are also consistent with those in [15]: TCP/ RED will become unstable when delay increases, or when link capacity increases.…”

“…In particular, the stability of TCP systems has been studied from the point of window-based flow control [6,[13][14][15][16][17]26,27] and rate control [19,21]. A linear model of a RED gateway with TCP connections was developed in [13,14] to characterize the stability region of the system and guideline for setting RED parameters was provided for the single-node analysis.…”

“…Notably, [6] proved the stability of TCP/RED systems without feedback delay, using a fluid model. The stability of TCP/RED systems with feedback delay has been studied in [15], which indicated that TCP/RED becomes unstable when the delay increases or, more strikingly, when link capacity increases. Our previous results provided some sufficient conditions to guarantee the stability of single-bottleneck AIMD/RED systems, considering feedback delays [26].…”

Bandwidth Problem in High Performance Packet Switching Network

“…Recently, reference [11] shows that the increase in the product of capacity and delay results instability in a random early discard [7], random early marking [6], proportional integral controller [9], and virtual queue [8]. In addition, it has been shown that an active queue management scheme can maintain stability over very high-capacity or large-delay links.…”

“…In choke method the warning is from the router, which has encountered congestion to the source station directly [4,11]. The intermediate nodes through which the packet has traveled are not warned.…”