A multiadaptive sampling technique for cost-effective network measurements

Abstract

The deployment of efficient measurement solutions to assist network management tasks without interfering with normal network operation assumes a prominent role in today's high-speed networks attending to the huge amounts of traffic involved. From a myriad of proposals for traffic measurement, sampling techniques are particularly relevant contributing effectively for this purpose as only a subset of the overall traffic volume is handled for processing, preserving ideally the correct estimation of network statistical behavior. In this context, this paper proposes MuST-a multiadaptive sampling technique based on linear prediction, aiming at reducing significantly the measurement overhead and still assuring that traffic samples reflect the statistical characteristics of the global network traffic under analysis. Conversely to current sampling techniques, MuST is a multi and self-adaptive technique as both the sample size and interval between samples are self-adjustable parameters according to the ongoing network activity and the accuracy of prediction achieved. The tests carried out demonstrate that the proposed sampling technique is able to achieve accurate network estimations with reduced overhead, using

Computational weight of network traffic sampling techniques

“…Despite the importance of sampling to reduce the computational effort of handling huge amount of data, most works in this area are only focused on analyzing the accuracy of traffic parameter estimation. The resource requirements of sampling techniques are covered through proposals which aims to reduce the memory usage [9] and the data volume involved in measurement process [2], maintaining the accuracy in traffic classification. Classic techniques [1] are also analyzed in terms of CPU load and memory usage in dedicated equipments [10], however there are no embracing studies analyzing and comparing the computational weight of existing sampling approaches, namely i.e., systematic, random and adaptive (see Section III), motivating the present work.…”

“…The sampling techniques evaluated correspond to the main techniques currently used in network measurement tools, i.e., systematic count-based, systematic time-based and random count-based [1]. In addition, two adaptive techniques are also evaluated, i.e., adaptive linear prediction [13] and multiadaptive sampling [2].…”

“…5) Multiadaptive sampling (MuST): this time-based technique resorts to a mechanism for identifying the level of network activity similarly to the adaptive linear prediction technique. However, the multiadaptive technique considers both the interval between samples and the sample size as adjustable parameters [2].…”

“…Conversely, in periods of less activity than predicted, in addition to sampling frequency reduction, the multiadaptive sampling also increases the sample size in order to acquire more information about the network without the risk of overloading the measurement point. Previous work has shown the ability of this technique in capturing correctly the network throughput with very low overhead as regards to the volume of data involved [2].…”

“…This study is however relevant as measurements should be carried out without compromising the performance of network elements when executing tasks involving end-user traffic. Thus, taking conventional sampling techniques commonly used in today's network devices [1] and recent adaptive sampling techniques [2] as reference, this paper is devoted to evaluate the computational burden of sampling regarding memory and CPU usage in presence of real network traffic. The volume of sampled traffic and the accuracy of these techniques in the estimation of parameters such as throughput and mean packet size are also analyzed.…”