Reliable Multicast Service Mapping in Network (Function) Virtualization
MetadataShow full item record
Network Virtualization and Software-defined Networking (SDN) provides an effective way to reduce network provider’s costs by allowing multiple Virtual Networks (VNs) to share the underlying physical infrastructure. A VN consists of virtual nodes and virtual links and can be mapped to the shared Substrate Network (SN) by mapping each virtual node to a physical node (servers/switches/storage), and mapping each virtual link to a physical path (and allocating the necessary bandwidth) in a SN. Network Functions Virtualization (NFV) leverages virtualization technology to implement network functions in software (called Virtualized Network Functions (VNFs)) that can run on standardized high volume servers/switches/storage in what we call Telecom Clouds (TCs). A set of network services can be provisioned through Service Function Chain (SFC) which consists of a set of VNFs interconnected by a number of virtual links. To deploy a service in NFV environment mainly includes two processes: (1) VNF placement: for each type of VNF, at least one TC need to be placed to provide the network function and multiple VNFs can share the resources on the same TC; (2) SFC Mapping/ (some paper call it chaining or routing): each VNF node need to be mapped onto a certain TC and links among VNF nodes need to be mapped onto physical links within/among TCs. Many big data applications, distributed file systems (e.g., Map-Reduce), point-to-multipoint real-time and interactive applications (e.g., video-conferencing and IPTV) prefer multicast communications in order to improve the utilization of the physical resources. Recently, designing efficient strategies to accommodate VNs/SFCs with multicast service traffic onto the shared SN has received a lot of attention in both academic and industry. In addition, Orthogonal Frequency Division Multiplexing (OFDM)-based Elastic Optical Networks (EONs) becomes a good candidate for SN since it has high bandwidth, low cost and flexible modulation selection and spectrum allocation. Furthermore, reliability is a critical requirement for multicast service since the failure of one virtual node/link can cause the malfunction of multiple nodes that receive multicasting data from/through it. In this thesis, we aim to design algorithms which can map VNs/SFCs with multicast traffic efficiently in order to decrease resource consumption and strategies for mapping multicast VNs/SFCs with high reliability. As a preliminary work, we have mainly focused on (1) VN mapping strategies for multicast services and reliability issue within the mapping; (2) VNF placement and SFC mapping for unicast services; (3) reliability issues within VNF placement and SFC mapping for multicast services. More specifically, first we have studied the VN mapping problem for multicast service in OFDM-based EONs, with the objective of minimizing spectrum consumption, while considering multiple modulation formats, spectrum continuity and spectrum conflict constraints. We have formulated a Mixed Integer Linear Programming (MILP) model to mathematically describe the problem and proposed two heuristic algorithms which are the Greedy algorithm and the Integrated Genetic and Simulated Annealing (IGSA) algorithm to efficiently solve the problem. Secondly, we have investigated the VN mapping for multicast services with max-min fairness of reliability over general IP networks. We have used a MILP model to formulate the problem and proposed the Reliability-Aware Genetic (RAG) algorithm to efficiently solve it. In addition, we have extended VN mapping for multicast services with max-min fairness of reliability from general IP networks to OFDM-based EONs by taking care of efficient modulation selection and spectrum allocation with spectrum conflict and continuity constraints. Thirdly, we have studied VNF placement and SFC mapping for unicast services, with the objective of minimizing both VNF placement and SFC mapping cost. We have proposed an efficient heuristic solution called Virtual Topology Design based VNF Placement and SFC Mapping (VTD-PM) algorithm to solve the problem. Lastly, we have investigated the VNF placement and SFC mapping for multicast services with max-min fairness of reliability over general IP networks. We have used a MILP model to formulate the problem and proposed the Reliability-Aware Self-Adaptive Genetic (RSG) algorithm to efficiently solve it. In addition, we have extended VNF placement and SFC mapping for multicast services with max-min fairness of reliability from general IP networks to OFDM-based EONs by taking care of efficient modulation selection and spectrum allocation with spectrum conflict and continuity constraints.