Large scale probabilistic available bandwidth estimation

Abstract

The common utilization-based definition of available bandwidth and many of the existing tools to estimate it suffer from several important weaknesses: i) most tools report a point estimate of average available bandwidth over a measurement interval and do not provide a confidence interval; ii) the commonly adopted models used to relate the available bandwidth metric to the measured data are invalid in almost all practical scenarios; iii) existing tools do not scale well and are not suited to the task of multi-path estimation in large-scale networks; iv) almost all tools use ad-hoc techniques to address measurement noise; and v) tools do not provide enough flexibility in terms of accuracy, overhead, latency and reliability to adapt to the requirements of various applications. In this paper we propose a new definition for available bandwidth and a novel framework that addresses these issues. We define probabilistic available bandwidth (PAB) as the largest input rate at which we can send a traffic flow along a path while achieving, with specified probability, an output rate that is almost as large as the input rate. PAB is expressed directly in terms of the measurable output rate and includes adjustable parameters that allow the user to adapt to different application requirements. Our probabilistic framework to estimate network-wide probabilistic available bandwidth is based on packet trains, Bayesian inference, factor graphs and active sampling. We deploy our tool on the PlanetLab network and our results show that we can obtain accurate estimates with a much smaller measurement overhead compared to existing approaches.

An end-to-end measurement and monitoring technique for the bottleneck link capacity and its available bandwidth

“…This latter part probe packets pattern is similar to the chirp-style, but in the reversed manner. [18] defines the probabilistic available bandwidth derived from the dynamics and variability of the available bandwidth, and try to estimate the large-scale probabilistic available bandwidth in a similar way to the tomography. It is based on the packet trains, Bayesian inference, factor graphs and active sampling.…”

“…Another version of the utilization from (18) can be defined as the mean utilization of a probe round averaged from the unit utilization of each packet-pair as follows.…”

“…The intrusion-based congestion inducing approaches [9][10][11][12][13][14][15]17,18] prevailing in the available bandwidth measurements inevitably have the drawbacks of affecting to the network resource usages and the flows that share the resources, and the situation becomes even worse when such approach requires iterative congestions to obtain just one valid measurement result. Such case wastes the network resources and requires large probe overhead, and so may not be highly practical for the monitoring purposes.…”

Available Bandwidth Estimation Using Collision Probability, Idle Period Synchronization and Random Waiting Time in MANETs: Cognitive Agent Based Approach

“…The accuracy of these techniques is still not satisfactory. The active techniques present in literature are Variable Packet Size (VPS) [11], SelfLoading Periodic Streams (SLoPS) [12,13], Train of Packet Pairs (TOPP) [14], Packet Train Pair (PTP) [15], SelfLoading Decreasing Rate Train (SLDRT) [16], Probabilistic Available Bandwidth (PAB) [17], Initial Gap Increasing (IGI) [18], Twoway Available Bandwidth Estimation (TWABE) [19], Gaps of Nonadjacent Probing packet (GNAPP) [20], Network Link Characteristics using PacketPair Dispersion (NLCPPD) [21], Feedbackassisted Robust Estimation of Available Bandwidth (bTRack) [22], Adaptive Available BE (AABE) [23], Bandwidth Available in RealTime (BART) [24], MultiRate BART (MRBART) [25], Minimal Backlogging Techniques (MiBT) [26], Distributed Admission Control For MANET Environment (DACME) [27], Reactive Bandwidth Measurement (RBM) [28,29], NeuroFuzzy Estimator (NFE) [30], Dualcarrier Sense with Parallel Transmission Awareness (DCSPT) [31], Packet Probing with RTS/CTS Handshake (PPRCH) [31]. In passive techniques, network statistic data is collected at the router level to infer network performance information from passive observation, which does not disturb the network traffic but waste limited computation and storage resources causing the delay.…”

Joint In-network Video Rate Adaptation and Measurement-Based Admission Control: Algorithm Design and Evaluation

“…Furthermore, several studies have provided a comparison of the different available tools, showing that they differ in accuracy and scalability [17,18]. More recently, Thouin et al [19] proposed a probabilistic available bandwidth estimation tool that links the maximum available bandwidth with the probability that the calculated bandwidth can be achieved. A taxonomy of common available bandwidth estimation tools has been presented by Strauss et al [20].…”