An experimental evaluation of the impact of heterogeneous scenarios and virtualization on the available bandwidth estimation tools

Abstract

Mobile Broadband (MBB) access networks are becoming more and more used worldwide, and the devices adopted to access them are increasing in number and complexity (smart-phones, mobile hotspots, vehicular infotainment systems). The highly dynamic nature of such scenarios calls for continuous monitoring and measurement of the network. To this aim, the Available Bandwidth is a performance metric of the utmost importance, albeit hard to estimate in uncontrolled scenarios. Shared experimental testbeds such as MONROE are becoming available to offer in-the-field MBB experimenting facilities. In this context, the SOMETIME project is focused on providing MONROE and similar testbeds with the tools to measure Available Bandwidth in MBB scenarios, also taking advantage of the benefit of SDN to perform active and passive measurements. In line with experimental activities planned by the project roadmap, in this paper we discuss the suitability of a number of publicly released ABw estimation tools when run in heterogeneous scenarios. Experimental results confirm that (i) the experimental scenario in which the tools run heavily impacts their performance in terms of accuracy; (ii) the entity of cross-traffic may have different effects on some of the tools, unacceptably undermining the estimation accuracy, depending on limitations of both the specific tools and the setup.