Class-based thresholds: Lightweight active router-queue management for multimedia networking.

摘要

In the best-effort Internet, congestion can cause severe degradations in the performance of both reliable data transfer flows and multimedia flows. Reliable data transfers are typically based on TCP, a responsive protocol that responds to congestion by reducing its transmission rate. In contrast, multimedia applications are typically based on unresponsive protocols, such as UDP, where the transmission rate is determined by the application independent of network conditions. There exists a fundamental tension between these protocols. During periods of congestion, unresponsive traffic maintains its load, forcing responsive traffic to reduce its load. Consequently, unresponsive traffic benefits from this behavior, consuming more than its fair share of the bandwidth, while responsive flows receive less than their fair share. In the extreme, congestion collapse is possible. This offers a disincentive for using responsive protocols.; Recent proposals have attempted to address this problem by identifying and severely constraining all unresponsive traffic. However, multimedia places bounds on minimum levels of throughput and maximum tolerable latency. These bounds may be difficult or impossible to meet with responsive protocols. As such, we argue that in addition to favoring responsive traffic, multimedia should not be unduly constrained, and must also be isolated from the effects of other unresponsive traffic. This dissertation presents and evaluates a novel algorithm to allocate network bandwidth by allocating buffer space in a router's queue: Class-Based Thresholds (CBT). A threshold limits the average queue occupancy by a given class. The ratio between these thresholds determines the ratio between the bandwidth available to each class on the outbound link.; CBT and other router queue management algorithms from the literature (FIFO, RED, and FRED) are implemented and evaluated. We explain the relationship between performance and configuration parameters using analysis for CBT and experimental data for other queue management schemes. We empirically demonstrate that CBT effectively isolates TCP while providing better-than-best-effort service for multimedia by comparing CBT's performance to the optimal performance for other algorithms. Finally, we show CBT provides better protection for TCP than RED and FIFO and better multimedia performance than RED, FIFO, and FRED.