Skip to main content

Hitachi

Corporate InformationResearch & Development

January 8, 2016

Report from Presenter

The Fourth International Conference on Communications, Computation, Networks and Technologies (INNOV 2015) was held in Barcelona, Spain, from Nov. 15 to Nov. 19, 2015. This conference is part of Softnet which was organized by International Academy, Research, and Industry Association (IARIA) and focuses on network and computation areas. Hitachi made a presentation on a system that can recover from both a small network failure and a catastrophic network failure, entitled "A System for Managing Transport-network Recovery according to Degree of Network Failure."


Fig. 1 Overview of inside-area protection and
end-to-end protection schemes

Enlarge

Lately, networks have become an indispensable service in our daily lives since various application services are provided by them. If a network is out of service, people's lives would be considerably affected. Therefore, the network should recover promptly when a failure occurs. In the conference, we presented a network failure recovery system according to the degree of network failure. In the system, a whole network is divided into multiple network (NW) areas and managed by a network-management server. The system has a three-step procedure for recovery that includes "an inside-area protection", "an end-to-end protection", and "an operation-plane change" schemes. An overview of the inside-area protection and the end-to-end protection schemes is shown in Figure 1. First, a network-management server configures inside-area protection paths between one edge PTN (PE1) and another edge PTN (PE2) for each current path in every area. These protection paths are used to recover from a small network failure in each area. Second, an end-to-end protection path between PE1 and PE2 for each current path is configured by the network-management server. Each end-to-end protection path is used to recover from a small failure such as a link failure between areas when the system is not recovered by the inside-area protection.


Fig. 2 Overview of operation-plane change scheme
Enlarge

An overview of the third procedure, operation-plane change scheme, is shown in Figure 2. The network-management server calculates multiple recovery paths by assuming all possible area-based network failures before starting network operations. For example, the number of possible combinations of areas is 256 (i.e., 28) if the network is divided into eight areas. It stores calculated recovery configurations by assigning a unique identifier (ID) for each network failure pattern. The pre-calculate recovery configurations are also stored in each packet transport node (PTN). When the network-management server detects network failures, it determines a failure pattern and sends a recovery ID to related PTNs. The PTNs changes path configurations according to the recover ID and the network failures are recovered. We have constructed a prototype system and evaluated it. As a result, the system could recover from both a small network failure and a catastrophic network failure.

Acknowledgments

Part of this research was done within research project O3(Open, Organic, Optima) and programs, "Research and Development on Virtualized Network Integration Technology," "Research and Development on Management Platform Technologies for Highly Reliable Cloud Services," and "Research and Development on Signaling Technologies of Network Configuration for Sustainable Environment" supported by MIC (The Japanese Ministry of Internal Affairs and Communications).

  • Page top