A neural network (NN) scheme is proposed for congestion control in an ATM switch with time priorities. It is shown that in a prioritised switch it is necessary to monitor the buffer to be controlled as well as buffers with higher priorities. Furthermore, it is shown that the NN scheme in a time prioritised switch gives lower cell loss and delay when compared to the conventional binary scheme.
References
-
-
1)
-
M. Garrett and A. Fernandez. (1994). ‘Variable bitrate video bandwidth trace using MPEG code’. Available FTP: thumper.bellcore.com, Directory: pub/vbr.video.trace .
-
2)
-
Yao-Ching Liu, , Douligeris, C.: `Static vs. adaptive feedback congestion controller for ATM network', Globecom'95, 13-17 November 1995, Singapore, p. 291–295.
-
3)
-
Available FTP: ftp-info3.informatik.uni-wuerzburg.deDirectory: pub/MPEG/traces.
-
4)
-
D. Rumelhart ,
G. Hinton ,
R. Williams
.
(1986)
Learning internal representation by error back-propagation, Parallel distributed processing: Exploration in microstructure of cognition.
-
5)
-
R. Jain
.
Congestion control and traffic management in ATM networks.
Comput. Netw. ISDN Syst.
,
13 ,
1723 -
1738
-
6)
-
Yao-Ching Liu ,
C. Douligeris
.
Rate regulator with feedback controller in ATM networks- A neural network approach.
IEEE J. Sel. Areas Commun.
,
2 ,
200 -
208
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19981467
Related content
content/journals/10.1049/el_19981467
pub_keyword,iet_inspecKeyword,pub_concept
6
6