Abstract

We describe the adoption of simulation techniques to test the operation of a new Self-Stabilizing Bridge Protocol [4; 24] and compare its perfonnance with that of an existing industry standard. Specifically, the experiments relate to the comparison of the IEEE 802.1 (D) Spanning Tree Algorithm [25; 26] as the industry standard, with a newly proposed Loop-Detect Packet based Self-Stabilizing Bridge Protocol. The language chosen for this work was Pascal Plus, which is an extension of Standard Pascal providing concurrent programming structures such as Processes and Monitors. The role of simulation in Cornmunications and Networks is highlighted and previous work surveyed. The problem under investigation relates to the use of devices, called Bridges, to connect two or more Local Area Networks (LANS) to form a larger interconnected system termed a Multi-LAN. Over recent years, reliability of communications systems has become an overwhelming concern to network managers. In order to increase reliability at the multi-LAN scale, it is usual to install parallel Bridges, thereby providing alternate paths between neighboring LANs. The inherent problem with such an approach is that it may lead to,more than one Bridge forwarding the same traffic to and from the adjacent LANs which they connect. This can cause duplicate packets, unnecessary traffic loading and, in some cases infinite looping of packets which severely degrade the performance of the communications channel they traverse. The IEEE 802.1 committee have defined an algorithm to address these problems, known as the Spanning Tree Algorithm. Although the underlying physical topology may be an arbitrary mesh, this scheme prunes the topology to a logical topology with no cycles. It does so by electing a single ROOT Bridge for the entire system, and subsequently electing a single Designated Bridge (active) for each LAN, with all other parallel Bridges assigned Backup status (passive). After the algorithm has converged, the Root is responsible for maintaining the logical tree, by periodically transmitting HELLO packets on all of its ports, which traverse the logical tree with the assistance of Designated Bridges. On receiving such HELLO packets, the Designated Bridges assume the logical topology is intact and forward the respective HELLO packets on their otihvard ports. However, if a Designated Bridge does not receive an HELLO packet within the defined timeout, it must assurne that the logical connectivity has been broken, and takes steps to regain status-quo. From a detailed investigation of the Spanning Tree Algorithm, several weaknesses in its design were discovered including the fact that it does not automatically make any attempt to limit the scope of maintenance traffic to those areas with cycles, and it is based on a centralized philosophy and as such, the failure of the Root forces the remaining Bridges to recornpute their slates from scratch. A new algorithm was designed by the Authors to overcome most of the problems with the existing standard. In particular, having pruned the physical tapology to a logical tree by use of special Loop Detect packets, the new algorithm places the responsibility of maintaining the logical connecfivity on the Backup (off) Bridges. The consequcence of this feature is that the Bridges concerned with maintaining the logical topology can fail, withont any dismption to the existing connectivity, as would be the case in the IEEE standard. By use of additional facilities the new algorithm also restricts the scope of maintenance traffic to those areas with cycles. Given the difficulties in using existing networks for protocol development (or any other software development for that matter), we opted for simulation in Pascal Plus as a means to test and validate the operation of our new protocol. To do this we adopted a layered approach to model conceptualization, which attempted to reflect the ISO OSI Reference Model. We present the simulation results and conclusions supporting the use of modelling techniques in the constantly evolving area of communications and networks which should assist others in assessing the practical issues involved.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.