Virtual network function placement for service function Chain in data center network | |
| Author | Hnin Pann Phyu |
| Call Number | AIT RSPR no.TC-16-05 |
| Subject(s) | Virtual computer systems Software-defined networking (Computer network technology) Data centers--Computer networks |
| Note | A research submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Telecommunication |
| Publisher | Asian Institute of Technology |
| Abstract | Because of significantly increasing number of various type of network services in today enterprise network, decreasing the time and cost of deploying this new network services become an important issue. Network Function Virtualization (NFV) in integration with software-defined networking concepts offers cost-efficient, scalable and more rapid deployment and flexible chaining of new network services. It allows allow shifting network functions (NF) from physical appliances to a pure software-based blocks which is called virtual network functions (VNFs). In this research paper, we define the problem of VNF placement in NFV infrastructure where different VNFs can share the processing resources of the same NFV node and multiple network function chains can share the same VNF type. Furthermore, as network requirements continue to grow in both scale and variety, and at the same time, it is necessary to add and upgrade middleboxes which lead to the telecommunication service provider (TSP) comes out with high Capital Expenditures (CAPEX) and Operational Expenditures (OPEX). Therefore, with the aim of reducing the overall cost of service providers, we formulate the Integer Linear Programming (ILP), for small number of VNF instances and also propose a random VNFs placement algorithm and cost efficient centrality-based VNFs placement (CCVP) algorithm for large scale networks. We intentionlly generate the results of random algorithm to proof the effectiveness of CCVP algorithm and linear program We use different real-world network topologies (Data Center, Internet 2 networks and RocketFuel ISP topologies) to verify that our ILP and CCVP algorithm is able to find the optimal number of VNF instances as well as able to reduce the overall cost of service providers effectively. According to our analysis results, we have found that CCVP algorithm is able to find a VNF deployment that satisfies the trade-off between the minimum number of VNF instances and the minimum bandwidth utilization. Finally, for the system validation, we compare our CCVP algorithm with the very recent and relevant algorithms; dynamic programming based heuristic (NFO-DP) and the binary search based heuristic (NFO-BS). |
| Year | 2016 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Teerapat Sanguankotchakorn; |
| Examination Committee(s) | Crespi, Noel ;Attaphongse Taparugssanagorn ;Nakamura, Shinichi ; |
| Scholarship Donor(s) | UNOCAL Myanmar Offshore Co.Ltd, Chevron; |
| Degree | Research Studies Project Report (M.Eng.) - Asian Institute of Technology, 2016 |