Block loss reduction in ATM networks

  • Authors:

  • Vijay Srinivasan;Anoop Ghanwani;Erol Gelenbe

  • Affiliations:

  • IBM Corporation, P.O. Box 12195, Research Triangle Park, NC 27709, USA;Department of Electrical Engineering, Duke University, Durham, NC 27708-0291, USA;Department of Electrical Engineering, Duke University, Durham, NC 27708-0291, USA

  • Venue:
  • Computer Communications
  • Year:
  • 1996

Quantified Score

Hi-index 0.24

Visualization

Abstract

Traditional approaches to guaranteeing Quality-of-Service (QoS) in ATM networks have focused on performance metrics such as cell loss probability, end-to-end cell delay and delay jitter. However, the block loss rate is a more meaningful metric for applications such as medical imaging and real-time video, and for high-level protocols such as IP which will use ATM as the transport mechanism. A block is defined as a group of consecutive ATM cells. For instance, a block may correspond to an IP datagram or to a macroblock within a medical image. A block loss occurs when a single cell from a block is lost. Consequently, we propose a technique for reducing block loss in ATM networks. Our method combines two well-studied approaches for minimizing the impact of information loss during transport over ATM networks which have been considered independently of each other in the past: priority-based cell discarding and forward error correction (FEC). We present a simple analysis which shows that our scheme could potentially reduce the block loss rate by several orders of magnitude for many cases of interest. We also formulate and solve a constrained optimization problem that can be used by a call admission control mechanism to determine optimal operating parameters. The analysis that we present is known to be somewhat optimistic since in reality cell losses in ATM-based networks are highly correlated. Consequently, we conduct extensive tests of the scheme via simulation. We also propose and study a buffer management algorithm which we call Adaptive Pushout (ADP) that accounts for correlations between cell losses. The ADP algorithm reduces block loss rates to nearoptimal levels. We consider further performance enhancements to the ADP policy that deliver excellent performance even under very heavy load. Practical implementation issues for the various schemes suggested in this paper are also addressed.