Open Access. Powered by Scholars. Published by Universities.®

Physical Sciences and Mathematics Commons

Open Access. Powered by Scholars. Published by Universities.®

Faculty of Informatics - Papers (Archive)

2000

Asynchronous transfer mode

Articles 1 - 2 of 2

Full-Text Articles in Physical Sciences and Mathematics

Label Switching Using The Ipv6 Address Hierarchy, P. Boustead, Joe F. Chicharo Nov 2000

Label Switching Using The Ipv6 Address Hierarchy, P. Boustead, Joe F. Chicharo

Faculty of Informatics - Papers (Archive)

Current label switching protocols can use routing, address, and address hierarchy information to group flows for cut-throughs that bypass IP forwarding. This paper examines a label switching solution that uses the IP version 6 (IPv6) address structure to classify and cut-through flows based on address hierarchy. The performance of this approach is examined using actual backbone traffic traces with associated hierarchical address information obtained from Internet address registries, routing arbiter databases and route servers. This hierarchical address information is used to map a hierarchical address structure over the packet level trace. We investigate the relationship between aggregation bit-mask size versus …


An Examination Of Ip/Atm Cut-Through Forwarding In Dynamically Routed Networks, P. Boustead, Joe F. Chicharo Aug 2000

An Examination Of Ip/Atm Cut-Through Forwarding In Dynamically Routed Networks, P. Boustead, Joe F. Chicharo

Faculty of Informatics - Papers (Archive)

Multiprotocol over ATM (MPOA), IP switching and multiprotocol label switching (MPLS) have distinctly different mechanisms for cut-through packet forwarding. MPOA and IP switching use flow-based cut-through (FBC) forwarding while MPLS uses routing table linked cut-through forwarding (TLC). This paper examines the sensitivity of each these cut-through forwarding mechanisms to changes in underlying routing tables. We examine a scenario where a congestion-sensitive dynamic routing protocol, such as OSPF optimised multipath, leads to frequently changing routing tables. We show that FBC forwarding reacts significantly worse than flow length distributions predict, taking up to 1200 seconds to react to route changes and forward …