Home // International Journal On Advances in Networks and Services, volume 8, numbers 1 and 2, 2015 // View article

A Network-disaster Recovery System using Multiple-backup Operation Planes

Authors:
Toshiaki Suzuki
Hideki Endo
Isao Shimokawa
Kenichi Sakamoto
Hidenori Inouchi
Taro Ogawa
Takanori Kato
Akihiko Takase

Keywords: network management; disaster recovery; packet transport; reliable network

Abstract:
A "network-disaster recovery system" using multiple-backup operation planes is proposed. Under this system, a whole network is separated into multiple areas. Before starting network operations, a network-management server calculates recovery paths for every possible failure in network area and distributes them with a recovery identifier (ID) for each network-area-failure pattern (on the backup operation plane). Network nodes receive and store the recovery IDs and recovery configurations. The network-management server determines a failure pattern after detecting the network-area failures and distributes the recovery ID to related network nodes. The network nodes that received the recovery ID start data transmission according to the path configurations specified by the recovery ID. After the completion of these procedures, the network-area failures are swiftly recovered. A prototype system (composed of a network-management server and 96 simulated packet-transport nodes) with a graphical viewer was implemented, and its performance was evaluated. According to the results of the evaluation, all recovery-path configurations for 1,000 pseudo-wires (PWs) (namely, transmitting the recovery ID to the related network nodes and using a recovery-path database specified by the ID) were done within 100 milliseconds after the network-area failures were detected. On the condition that the configuration time depends on the size of the recovery-path database, the proposed system takes about one minute and 40 seconds in the case of 1,000,000 PWs. On the other hand, a restoration scheme under the same evaluation conditions used for the proposed system takes over 10 minutes to recalculate recovery paths from detection of the first area-based network failure. That is, the proposed recovery scheme can recover network-area failures faster than the conventional restoration scheme can.

Pages: 118 to 129

Copyright: Copyright (c) to authors, 2015. Used with permission.

Publication date: June 30, 2015

Published in: journal

ISSN: 1942-2644