<p>The IP multimedia subsystem (IMS) call session control function (CSCF) has been recognized as the core control platform for the next-generation network (NGN). Our previous study presented an NFV-based virtualized IMS CSCF denoted as MISR vCSCF. This paper advocates innovative research by systematically developing a refactoring process to rearchitect the IMS CSCF. First, we decomposed the Serving CSCF (S-CSCF) and interrogation CSCF (I-CSCF) into smaller network functions. Then, we designed two quantitative indicators, message exchange reduction (MER) and scaling side effect (SSE), to evaluate the impacts of merging network functions. To select an optimal combination of merging some network functions to minimize the queuing delay under the restricted resource cost, we proposed a Pareto-Optimal heuristic algorithm with MER and SSE to refactor the IMS CSCF and derive the Pareto-Optimal Refactored IMS CSCFs (POR-IMS vCSCFs). In addition, we used a brute-force approach to find the optimal solutions by enumerating all possible solutions, called the BF-based Refactored CSCF(BFR-IMS vCSCF). The BFR-IMS solutions are a subset of the POR-IMS solutions. Moreover, the POR-IMS solutions had shorter queueing delays and better resource utilization rates than the MISR vCSCF.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

A pareto-optimal heuristic algorithm to refactor IMS core for cloud-native NFV

  • Wei-Kuo Chiang,
  • Yuan-Hsiang Chih

摘要

The IP multimedia subsystem (IMS) call session control function (CSCF) has been recognized as the core control platform for the next-generation network (NGN). Our previous study presented an NFV-based virtualized IMS CSCF denoted as MISR vCSCF. This paper advocates innovative research by systematically developing a refactoring process to rearchitect the IMS CSCF. First, we decomposed the Serving CSCF (S-CSCF) and interrogation CSCF (I-CSCF) into smaller network functions. Then, we designed two quantitative indicators, message exchange reduction (MER) and scaling side effect (SSE), to evaluate the impacts of merging network functions. To select an optimal combination of merging some network functions to minimize the queuing delay under the restricted resource cost, we proposed a Pareto-Optimal heuristic algorithm with MER and SSE to refactor the IMS CSCF and derive the Pareto-Optimal Refactored IMS CSCFs (POR-IMS vCSCFs). In addition, we used a brute-force approach to find the optimal solutions by enumerating all possible solutions, called the BF-based Refactored CSCF(BFR-IMS vCSCF). The BFR-IMS solutions are a subset of the POR-IMS solutions. Moreover, the POR-IMS solutions had shorter queueing delays and better resource utilization rates than the MISR vCSCF.