TY - JOUR
T1 - A Virtual Network embedding approach to preserving node closeness
AU - Liu, Xin Gang
AU - Huai, Jin Peng
AU - Gao, Qing Yi
AU - Wo, Tian Yu
PY - 2012/12
Y1 - 2012/12
N2 - Virtual Network (VN) embedding is one of the key problems of network virtualization. It can be separated into two basic phases, the node mapping and the link mapping phases. Previous research mainly focused on the link mapping phase, while there are only few simple and crude solutions in node mapping phase. However, the result of node mapping has a significant impact on the performance of virtual network embedding. There are two major drawbacks of previous node mapping solutions. Firstly, nodes with sufficient bandwidth resources are first selected while finding the hosts, which is not good for resource utilization usually. Secondly, existing methods separate the two phases too clearly so that neighbor nodes in the virtual network may be mapped to faraway nodes in the substrate network, which causes virtual links to take more bandwidth resources. This paper proposes a new virtual network embedding algorithm with node distributed closely. We take all the substrate nodes that meet a virtual node's requirement as its candidate hosts, so that there are more choices while selecting its host. It is conducive to rational allocation of resources. After that, we choose those closely distributed substrate nodes as the hosts of the virtual nodes, and those adjacent nodes in the virtual network request are mapped to substrate nodes in neighborhood. As a result, fewer resources are needed to finish link mapping. Simulation experiments show that the proposed algorithm improves resource utilization and increases the acceptance ratio significantly.
AB - Virtual Network (VN) embedding is one of the key problems of network virtualization. It can be separated into two basic phases, the node mapping and the link mapping phases. Previous research mainly focused on the link mapping phase, while there are only few simple and crude solutions in node mapping phase. However, the result of node mapping has a significant impact on the performance of virtual network embedding. There are two major drawbacks of previous node mapping solutions. Firstly, nodes with sufficient bandwidth resources are first selected while finding the hosts, which is not good for resource utilization usually. Secondly, existing methods separate the two phases too clearly so that neighbor nodes in the virtual network may be mapped to faraway nodes in the substrate network, which causes virtual links to take more bandwidth resources. This paper proposes a new virtual network embedding algorithm with node distributed closely. We take all the substrate nodes that meet a virtual node's requirement as its candidate hosts, so that there are more choices while selecting its host. It is conducive to rational allocation of resources. After that, we choose those closely distributed substrate nodes as the hosts of the virtual nodes, and those adjacent nodes in the virtual network request are mapped to substrate nodes in neighborhood. As a result, fewer resources are needed to finish link mapping. Simulation experiments show that the proposed algorithm improves resource utilization and increases the acceptance ratio significantly.
KW - Compact distribution
KW - Optimization
KW - Virtual network
KW - Virtual network embedding
UR - https://www.scopus.com/pages/publications/84874391315
U2 - 10.3724/SP.J.1016.2012.02492
DO - 10.3724/SP.J.1016.2012.02492
M3 - 文章
AN - SCOPUS:84874391315
SN - 0254-4164
VL - 35
SP - 2492
EP - 2504
JO - Jisuanji Xuebao/Chinese Journal of Computers
JF - Jisuanji Xuebao/Chinese Journal of Computers
IS - 12
ER -