Research On Geographic Based Survivable Virtual Network Mapping Technologies

With the popularity of the Internet, a wide variety of Internet-based applications emerge in endlessly, these new applications present different demand on Internet, so that the existing Internet architecture is facing great challenges, the development of the Internet shows fossilization. In this case, Network Virtualization(NV) was proposed, in the current, NV has been considered by everyone an effective and feasible solution to the ossified development of the current Internet. However, in order to apply this solution, a major challenge facing first is how to map reasonable and effective the Virtual Network(VN) to the Substrate Network(SN), this is the Virtual Network Mapping Problem(VNMP). In the case of the limited physical resources, effective mapping of VN can not only improve the utilization of physical resources, reduce energy consumption, but also save cost of the user, which is very favorable, for the user and the Infrastructure Provider s(InPs); In addition, it is possible that the SN occurs failure, in order to ensure the survival of virtual nodes and virtual links, virtual network survival mapping is one of the key issues in the VNMP.This paper studies the three virtual network mapping issues:(1) in the case of considering the geographic of virtual nodes and that at most one node of the SN occur failure at a moment, how to survivable map a single VN(2) in the case of considering the geographic of virtual nodes and that at most one node of the SN occur failure at a moment, how to survivable map the multiple VNs which arrive simultaneously(3) in the case of considering the geographic of virtual nodes and that the SN occur regional failure at a moment, how to survivable map a single VN, as far as possible, which tries to guide that these virtual nodes which can be mapped to the same physical node are mapped to the same physical node, so that, the resources between the backup nodes can be shared, reducing the costs of mapping virtual network’s backup node.For the first question, this paper proposes G-SVNM( Geographic based Survivable VN Mapping)algorithm and GG-SVNM(Geographic-Guided Survivable VN Mapping) algorithm. G-SVNM algorithm considers the impact of virtual nodes’ geographic on virtual network mapping; GG-SVNM algorithm proposes that the geographic of virtual nodes can guide the mapping of VN ’s backup nodes.For the second question, this paper proposes GG-SMVNM(Geographic-Guided Survivable Multiple VN Mapping)algorithm. In the GG-SMVNM algorithm, the mapping of backup nodes and backup links not only consider the resource sharing in each VN, but also consider the case of a resource-sharing between these simultaneous arrival virtual networks, which saves the physical network resources, making the mapping cost less.For the third question, this paper proposes G-IRVNM( Geographic based Incremental Recovery VN Mapping)algorithm and G-FRVNM(Geographic based Full Recovery VN Mapping) algorithm. G-IRVNM algorithm only recovery the failed virtual nodes and virtual links, don’t migrate the not-failed virtual nodes and virtual links; Once failure occurs, G-FRVNM algorithm migrate all the virtual nodes and virtual links, which tries to recovery the failed virtual nodes and virtual links.