Effect of bursty source traffic on rate-based ABR congestion control schemes

Congestion control is an important function of traffic management in Asynchronous Transfer Mode (ATM) net works. Available Bit Rate (ABR) is a category of service for ATM networks wherein the source adapts its generation rate to network conditions based on feedback. In this paper we study closed-loo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hegde, M., Moh, W.M.
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Congestion control is an important function of traffic management in Asynchronous Transfer Mode (ATM) net works. Available Bit Rate (ABR) is a category of service for ATM networks wherein the source adapts its generation rate to network conditions based on feedback. In this paper we study closed-loop ABR congestion control using special cells called Resource Management (RM) cells to relay feedback to the source. In particular, we investigate both the effectiveness of rate-based ABR congestion control in the presence of bursty source traffic and the relationship between the burst time scale and the ABR control time scale. Two ABR congestion control schemes, the ABR Explicit Forward Congestion Indication (EFCT) and ABR Congestion Indication (CT) schemes, are compared with Unspecified Bit Rate (UBR) transport which makes no effort to control congestion. Traffic sources of various burst lengths of 100, 1000, 10000, and an equal mix of 100 and 10000 cells are used in simulations. It is found that ABR congestion control effectively controls low frequency, medium to long term traffic load transients, but does not control high frequency, short term load transients. In the latter case, ABR control is not necessary since short term transients do not require large amount of buffering. Of the two AABR schemes considered, the more sophisticated ABR CT scheme performs significantly better than the ABR EFCI scheme in terms of delay and buffer occupancy.
DOI:10.1109/ICBC.1996.887789