Presentation 2016-11-25
Dynamic Resource Migration in Virtualized Networks by Using Reinforcement Learning
Takaya Miyazawa, Ved P. Kafle, Hiroaki Harai,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In network environments where multiple virtual network operators (VNOs) create VNs over common edge, core, and datacenter networks, the physical resources should be dynamically and agilely reallocated and migrated among VNs in various situations, such as to construct a new urgent VN in case of occurrence of some urgent contingency, or to satisfy quality of service (QoS) requirements of non-urgent VNs in case of time-varying network traffic conditions as much as possible. In this paper, we propose methods to automatically and dynamically select and migrate resources of non-urgent VNs after preferentially allocating necessary resources to the urgent VN. In particular, our proposal applies reinforcement learning for selection of resources from alternate places to satisfy the QoS requirements. Simulation results show that the proposed dynamic resource migration method can make it easier to satisfy non-urgent VNs’ QoS requirements compared to static resource assignment methods. Moreover, we show that, applying reinforcement learning can increase the number of times QoS requirements are satisfied for VNs in comparison with the dynamic method with completely random resource selection.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Virtual Network / Resource Migration / Reinforcement Learning / Quality of Service
Paper # NS2016-115
Date of Issue 2016-11-17 (NS)

Conference Information
Committee NS / CQ / ICM
Conference Date 2016/11/24(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Shimonoseki Chember of Commerce and Industry
Topics (in Japanese) (See Japanese page)
Topics (in English) Network Quality, Network Measurement and Management, Network Virtualization, Network Service, General
Chair Hideki Tode(Osaka Pref. Univ.) / Kyoko Yamori(Asahi Univ.) / Toshio Tonouchi(NEC)
Vice Chair Yoshikatsu Okazaki(NTT) / Takanori Hayashi(NTT) / Hideyuki Shimonishi(NEC) / Manabu Nakagawa(NTT Communications) / Yuji Nomura(Fujitsu Labs.)
Secretary Yoshikatsu Okazaki(Kyushu Inst. of Tech.) / Takanori Hayashi(NTT) / Hideyuki Shimonishi(Osaka Univ.) / Manabu Nakagawa(Keio Univ.) / Yuji Nomura(NEC)
Assistant Shohei Kamamura(NTT) / Hirantha Abeysekera(NTT) / Norihiro Fukumoto(KDDI R&D Labs.) / Masaharu Hattori(KDDI R&D Labs.)

Paper Information
Registration To Technical Committee on Network Systems / Technical Committee on Communication Quality / Technical Committee on Information and Communication Management
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Dynamic Resource Migration in Virtualized Networks by Using Reinforcement Learning
Sub Title (in English)
Keyword(1) Virtual Network
Keyword(2) Resource Migration
Keyword(3) Reinforcement Learning
Keyword(4) Quality of Service
1st Author's Name Takaya Miyazawa
1st Author's Affiliation National Institute of Information and Communications Technology(NICT)
2nd Author's Name Ved P. Kafle
2nd Author's Affiliation National Institute of Information and Communications Technology(NICT)
3rd Author's Name Hiroaki Harai
3rd Author's Affiliation National Institute of Information and Communications Technology(NICT)
Date 2016-11-25
Paper # NS2016-115
Volume (vol) vol.116
Number (no) NS-322
Page pp.pp.79-84(NS),
#Pages 6
Date of Issue 2016-11-17 (NS)