
It is essential for the Future Internet to fully support multihoming and select most appropriate paths for Concurrent Multipath Transfer (CMT). In real complex networks, different paths are likely to overlap each other and even share bottlenecks which can weaken the path diversity gained through CMT. Spurred by this observation, it is necessary to select multiple independent paths insofar as possible. However, the path correlation lurks behind the IP/network layer topology, so we have to fall back to end-to-end probes to estimate this correlation by analyzing path delay characteristics. In this paper, we present the first step towards a new topic of correlation-aware multipath selection, with formal and systematic problem definition, modeling and solution. Based on a well-designed delay probing, a Grouping-based Multipath Selection (GMS) mechanism is developed to avoid underlying shared bottlenecks between topologically joint paths. In addition, we further propose a practical functionality framework and define a novel multihoming sublayer for the exchange of the multipath capabilities. Extensive simulations demonstrate that the GMS under different network conditions performs much better than other selection schemes, even with burst background traffic.
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 30 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Top 10% | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Top 10% |
