No abstract

Abstract. In this paper we propose the concept of an end-to-end (e2e) SelfProtecting Multi-Path (SPM) as a protection switching mechanism that may be implemented, e.g., in Multiprotocol Label Switching (MPLS) networks. In case of local outages, resilient networks redirect the traffic from a failed link over an e2e backup path to its destination. In this case, Quality of Service (QoS) can only be provided if sufficient extra capacity is available. If backup capacity can be shared among different backup paths, multi-path routing allows for considerable savings regarding this extra capacity. The SPM consists of disjoint paths that carry the traffic bothin normal operation mode and during local outages. If a partial path is affected by a network failure, the traffic is just distributed to the remaining working paths. This structure is easy to configure and the switching to failure mode operation is simple since no signalling is required. Based on analytical results, we show that Ioad balancing of the traffic across the disjoint paths can reduce the required backup capacity significantly. The backup performance depends strongly on the network topology, and the SPM outperforms simple Open Shortest Path First (OSPF) rerouting by far.

<p class="MsoNormal" style="text-align: left; margin: 0cm 0cm 0pt;" align="left"><span class="text"><span style="font-family: ">This article investigates the transient behavior of experience-based admission control (EBAC) in case of traffic changes. EBAC is a robust and resource-efficient admission control (AC) mechanism used for reservation overbooking of link capacities in packet-based networks. Recent analyses gave a proof of concept for EBAC and showed its efficiency and robustness through steady state simulation on a single link carrying traffic with constant properties. The contribution of this paper is an examination of the memory from which EBAC gains its experience and which strongly influences the behavior of EBAC in stationary and nonstationary state. For the latter, we investigate the transient behavior of the EBAC mechanism through simulation of strong traffic changes which are characterized by either a sudden decrease or increase of the traffic intensity. Our results show that the transient behavior of EBAC partly depends on its tunable memory and that it copes well with even strongly changing traffic characteristics.</span></span></p>

Classical admission control approaches take either descriptor or measurement based information about the traffic into account without relating them to each other. We propose a experience-based AC (EBAC) which uses an empirical percentile of the effective reservation utilization to determine a suitable overbooking factor. In this paper, we show the impact of different measurement time scale resolutions and different quantiles on the performance of the system. We propose aging mechanisms for statistic collection to make the system adaptive to traffic mixes that change over time. We illustrate their effectiveness by simulation results.