Designing a Predictable Internet Backbone with Valiant Load-Balancing

Abstract

Abstract. Network operators would like their network to support current and future traffic matrices, even when links and routers fail. Not surprisingly, no backbone network can do this today: It is hard to accurately measure the current matrix, and harder still to predict future ones. Even if the matrices are known, how do we know a network will support them, particularly under failures? As a result, today's networks are designed in a somewhat ad-hoc fashion, using rules-of-thumb and crude estimates of current and future traffic.Previously we proposed the use of Valiant Load-balancing (VLB) for backbone design. It can guarantee 100% throughput to any traffic matrix, even under link and router failures. Our initial work was limited to homogeneous backbones in which routers had the same capacity. In this paper we extend our results in two ways: First, we show that the same qualities of service (guaranteed support of any traffic matrix with or without failure) can be achieved in a realistic heterogeneous backbone network; and second, we show that VLB is optimal, in the sense that the capacity required by VLB is very close to the lower bound of total capacity needed by any architecture in order to support all traffic matrices.

Comparative Study of Ip Backbone Network Modeled Optimally and Through Bandwidth Overprovision

“…Traffic modelling is the most contentious and key element in the design and simulating of communication networks [4], [5]. A clear understanding of the nature of traffic in the target system and subsequent selection of an appropriate random traffic model are critical to the success of design of the system [4].…”

“…Bandwidth over-provisioning is commonly used to protect network performance against traffic variations. It typically involves dimensioning links so that their bandwidth exceeds the expected traffic load by large margins, to ensure that the link can absorb both expected and unexpected traffic fluctuations [5], [8]. A major reason for use of the bandwidth overprovision method is the apparent difficulty involved in the estimation of the traffic matrix [5], [8].…”

“…It typically involves dimensioning links so that their bandwidth exceeds the expected traffic load by large margins, to ensure that the link can absorb both expected and unexpected traffic fluctuations [5], [8]. A major reason for use of the bandwidth overprovision method is the apparent difficulty involved in the estimation of the traffic matrix [5], [8]. This, more than anything else, has raised serious doubts about this traditional approach as a veritable choice for the optimum design of IP backbone networks as traffic is fundamental to the design of any telecommunications network.…”

“…The main challenges that service providers face are to reduce costs, and to enable new value-added services that were previously not possible [5], [13]. The traffic explosion in the backbone network comes in unpredictable volumes, characteristics, and patterns [13].…”

ProActive routing in scalable data centers with PARIS

“…In a full-mesh network of N nodes, where a node can send traffic at a maximum rate r to another node, the link capacity of mesh links required to support any traffic matrix using VLB is only r( 2 N ) [33]. For implementing VLB in our topology, the link capacity required between any pair of core switches will be r( 4 N ), where in the first phase a packet entering the core layer is forwarded to an APS after bouncing it off a randomly selected core switch, and in the second phase it is forwarded to an egress core switch from the APS after again bouncing it off a random core switch.…”