Research Of Network Slicing Resource Management In Space-Air-Ground Integrated Networks Based On Service Function Chain Orchestration

The space-air-ground integrated network plays a vital role in achieving global seamless coverage and providing ubiquitous communication services.The integration of the space network,airborne networks,and ground network,along with leveraging the strengths of various network types,enables the spaceair-ground integrated network to offer users comprehensive and diverse communication services.However,the highly heterogeneous space-air-ground integrated network involves different types of terminals,nodes,and resources,which increases the complexity of integrating the networks described above.Furthermore,the network offers various services,each of which is with distinct quality requirements and network characteristics.How to ensure the diversity and differentiation of these services also a crucial problem.Therefore,the introduction of network slicing technology is imperative to the space-air-ground integrated network.Network slicing masks the differences of the underlying physical networks,divides the network into multiple independent virtual networks through virtualization technology,and customizes and optimizes each virtual network to suit the service requirements.Each slice virtual network comprises different virtual network functions(VNFs),by combining multiple VNFs into a chain structure to form a service function chain(SFC)and assigning corresponding virtual resources to them,resources can be flexibly allocated and services can be customized.However,a key challenge brought by the introduction of network slicing technology is how to devise effective orchestration strategies to efficiently manage and schedule network slicing resources,achieving optimal match of corresponding services and resources.The complex architecture of the space-air-ground integrated network,diverse business types,and dynamic network characteristics further increase the difficulty of resource management in network slicing.To this end,this article focuses on the network slicing resource management in space-air-ground integrated network based on SFC orchestration.Effective SFC orchestration strategies are devised to achieve efficient resource management for different levels and scenarios of air-ground-space network slicing.The specific research contents and contributions are as follows:1.Research on slice-based SFC embedding strategy for network slice deploymentThis research focuses on achieving end-to-end space-air-ground integrated network slice deployment by embedding the SFCs of different slice service flows(referred to as slice-based SFC embedding),under the soft slice isolation mode,and proposes the corresponding slice-based SFC embedding strategies.When implementing slice-based SFC embedding,a series of problems caused by the introduction of the slice dimension are fully considered,such as diversified and differentiated requirements of different slice service flows,inter-slice and intra-slice VNF sharing,slice priority-aware service flow admission control and VNF placement restrictions,etc.The slice-based SFC embedding problem for network slice deployment is modeled as an optimization problem of maximizing service flow acceptance ratio and minimizing network resource overhead,and it is proved to be NP-hard.Considering the problem is difficult to solve the optimal solution when its scale is large,two heuristic slice SFC embedding algorithms in polynomial time are designed to solve the approximate optimal solution of the problem in a short running time.The simulation results verify that the proposed algorithms outperform the benchmark algorithm in different complex network topologies.Additionally,they achieve the slice flow acceptance ratio sum exceeding 0.99 times and total network resource cost within 1.22 times the optimal solution,respectively.2.Research on SFC embedding strategy for dynamic topologyThis research focuses on realizing the mapping of slice resources to services by embedding SFC,and proposes the corresponding embedding strategies of SFC,considering the dynamic topology of the space-air-ground integrated network.To characterize the dynamic nature of the network topology,time evolving graph is introduced to connect the quasi-static network topology during different time slots,and the mapping from slice resources to service flows is realized by SFC embedding.The admission control,SFC constraint,and received time slot decision of the service flow are considered comprehensively in SFC embedding.The dynamic topology-oriented SFC embedding problem is modeled as an optimization problem that maximizes service flow acceptance ratio and minimizes slice resource consumption.Considering the problem is NP-hard,an optimal algorithm based on column generation is designed to effectively get the solution of the problem.However,obtaining the optimal solution through the algorithm still costs a significant amount of time when the problem size is large.Therefore,a heuristic SFC embedding algorithm in polynomial time is designed to obtain an approximate optimal solution in a shorter running time.Considering the characteristic of the designed heuristic algorithm,an accelerated optimal algorithm is proposed by combining the column generation algorithm with the heuristic algorithm to speed up the process of obtaining the optimal solution.Simulation results demonstrate that the proposed heuristic algorithm is superior to the benchmark algorithm,and can achieve the flow acceptance ratio exceeding 0.96 times and total network resource consumption within 1.004 times the optimal solution,respectively.Meanwhile,the proposed accelerated optimal algorithm can effectively accelerate the solving process of the optimal solution.3.Research on multicast SFC orchestration strategy with user fluidity awarenessThis research focuses on user fluidity-aware multicast SFC orchestration in space-air-ground multicast service slicing,and proposes corresponding multicast SFC orchestration strategies.To adapt to the user fluidity(i.e.,frequent user arrival and departure)in multicast services,in addition to studying typical multicast SFC embedding,two new multicast SFC orchestration types are proposed,namely multicast SFC readjustment and multicast SFC expanding.They adjust and extend the corresponding slice resource configuration to maintain high resource utilization and user acceptance ratio.Furthermore,several key optimization factors are simultaneously considered for each kind of multicast SFC orchestration,such as flexible replication multicast packets,network resource capacity constraints,and source-destination delay constraint.The three multicast SFC orchestration problems are modeled as optimization problems that aim to maximize slice net revenue and proven to be NP-hard.Considering that obtaining the optimal solutions are challenging when the problems’sizes are large,three polynomial-time heuristic multicast SFC orchestration algorithms are designed to respectively solve the three problems,which can obtain approximate optimal solutions in a shorter running time.The simulation results demonstrate that the proposed three algorithms outperform the benchmark algorithms,and achieve the slice net revenue exceeding 0.96 times,0.75 times,and 0.84 times the optimal solutions,respectively.