Topology design and bandwidth allocation in ATM nets

Abstract

In emerging network technologies designed to support a variety of services, it is common to find that the packet switching service is implemented on top of a facility network. For example, in typical narrowband ISDN architectures, the packet switches installed a t some of the central offices (CO's) a r e interconnected with trunks derived from an underlying channelized facility network. Likewise, in future broadband ISDN's, the ATM switches will he connected with trunks obtained from a n underlying "pool" of fiber facilities interconnected by digital cross connect systems (DCS). In this paper, we address the design of a P/S network embedded into a backbone facility network. We formulate the problem as a network optimization problem where a congestion measure based on the average packet delay is minimized, subject to capacity constraints posed by the underlying facility trunks. The variables in this problem a r e the routing on the "express" pipes (i.e., the channels that interconnect the PIS modes), and the allocation of bandwidth to such pipes. We present a n efficient algorithm for the solution of the above problem and apply it to some representative examples. We show that for some test cases, the congestion measure is substantially reduced with respect to the values obtained when the embedded topology is kept identical to the hackbone topology. We also discuss dynamic reconfiguration schemes where the embedded topology is periodically adjusted to track the fluctuations in traffic requirements.

Off-line configuration of a MPLS over WDM network under time-varying offered traffic

“…Our formulation follows the multicommodity flows structure that was already adopted in several previous works (e.g. [1] [3] [5] [6]). A further novelty of this paper is that our MILP formulation considers a more complex cost function, which explicitely allows for explicit control over the balancing between opticallevel and packet-level resources.…”

“…In other words, a two layer matching must be found: i) between the traffic matrix and the logical topology at the upper layer and ii) between the logical topology and the physical topology at the lower layer. Similarly to [6], in our approach we assume that for the given physical topology (nodes, links) the sets of all possible candidate arcs from node s to d, hereafter denoted by A s,d , is pre-computed. Eventually, some space-reduction heuristic can be applied at this early step (see discussion in section IV) to cut-away from A s,d those arcs which are less likely to be selected in the optimal solution, in order to reduce the problem size and then the resolution time.…”

Cycle‐based algorithms for multicommodity network flow problems with separable piecewise convex costs

“…The problem can also be modelled as a mixed-integer nonlinear program [18], and can be seen as a particular case of the capacity and flow assignment problem (CFA), introduced by Gerla and Kleinrock [9]. To solve this problem, researchers have either tried to decouple the difficulty by alternately solving the capacity and the flow assignment problems (the so-called CA-FA procedure [6,7,10]), or used exact approaches like Benders decomposition [16].…”

High Level B-ISDN/ATM Traffic Management in Real Time

“…Besides the general appeal of a structured approach to tra c management, some of the motives behind VPs and VNs 2, 11 16,21,33,34,41,42 are: Reduced costs resulting from simpli ed transit exchanges; Faster call handling by excluding intermediate node processing at set-up time; Improved tra c management capabilities such as possibilities to redirect tra c in a congested or faulty network; A means for providing customerdedicated, closed VNs. For B-ISDN type networks, we may additionally gain simpli ed statistical multiplexing and grade of service control by grouping services into service classes SCs according to their characteristics, e.g.…”