Machine Learning for Network Slicing Resource Management:A Comprehensive Survey

doi:10.12142/ZTECOM.201904005

Abstract

Abstract:

The emerging technology of multi-tenancy network slicing is considered as an essential feature of 5G cellular networks. It provides network slices as a new type of public cloud services and therewith increases the service flexibility and enhances the network resource efficiency. Meanwhile, it raises new challenges of network resource management. A number of various methods have been proposed over the recent past years, in which machine learning and artificial intelligence techniques are widely deployed. In this article, we provide a survey to existing approaches of network slicing resource management, with a highlight on the roles played by machine learning in them.

Key words: 5G, machine learning, multi-tenancy, network slicing, resource management

HAN Bin, Hans D. SCHOTTEN. Machine Learning for Network Slicing Resource Management:A Comprehensive Survey[J]. ZTE Communications, 2019, 17(4): 27-32.

Figures/Tables 3

References 28

1	AllianceNGMN. Description of Network Slicing Concept [R]. 2016
2	ITU⁃R. Minimum Requirements Related to Technical Performance for IMT⁃2020 Radio Interface(s) [R]. Report ITU⁃R M.2410-0, 2017
3	SCIANCALEPORE V, CIRILLO F, COSTA⁃PEREZ X. Slice as a Service (SlaaS): Optimal IoT Slice Resources Orchestration [C]//2017 IEEE Global Communications Conference (GLOBECOM). Singapore, Singapore, 2017: 1–7. DOI: 10.1109/GLOCOM.2017.8254529 DOI
4	SALLENT O, PÉREZ⁃ROMERO J, FERRÚS R, et al. On Radio Access Network Slicing from a Radio Resource Management Perspective [J]. IEEE Wireless Communications, 2017, 24(5): 166–174. DOI: 10.1109/MWC.2017.1600220WC DOI
5	CABALLERO P, BANCHS A, DE VECIANA G, et al. Network Slicing for Guaranteed Rate Services: Admission Control and Resource Allocation Games [J]. IEEE Transactions on Wireless Communications, 2018, 17(10): 6419–6432. DOI: 10.1109/TWC.2018.2859918 DOI
6	WU D, ZHANG Z, WU S, et al. Biologically Inspired Resource Allocation for Network Slices in 5G⁃Enabled Internet of Things [J]. IEEE Internet of Things Journal, 2018. DOI: 10.1109/JIOT.2018.2888543 DOI
7	HAN B, JI L, SCHOTTEN H D. Slice as an Evolutionary Service: Genetic Optimization for Inter⁃Slice Resource Management in 5G Networks [J]. IEEE Access, 2018, 6: 33137–33147. DOI: 10.1109/ACCESS.2018.2846543 DOI
8	LI R, ZHAO Z, SUN Q, et al. Deep Reinforcement Learning for Resource Management in Network Slicing [J]. IEEE Access, 2018, 6: 74429–74441. DOI: 10.1109/ACCESS.2018.2881964 DOI
9	HAN B, FENG D, SCHOTTEN H D. A Markov Model of Slice Admission Control [J]. IEEE Networking Letters, 2019, 1(1): 2–5. DOI: 10.1109/LNET.2018.2873978 DOI
10	VO P L, NGUYEN M N H, LE T A, et al. Slicing the Edge: Resource Allocation for RAN Network Slicing [J]. IEEE Wireless Communications Letters, 2018, 7(6): 970–973. DOI: 10.1109/LWC.2018.2842189 DOI
11	D’ORO S, GALLUCCIO L, MERTIKOPOULOS P, et al. Auction⁃Based Resource Allocation in OpenFlow Multi⁃Tenant Networks [J]. Computer Networks, 2017, 115: 29–41. DOI: 10.1016/j.comnet.2017.01.010 DOI
12	BEGA D, GRAMAGLIA M, BANCHS A, et al. Optimising 5G Infrastructure Markets: The Business of Network Slicing [C]//2017 IEEE Conference on Computer Communications (INFOCOM). Atlanta, USA, 2017: 1–9. DOI: 10.1109/INFOCOM.2017.8057045 DOI
13	CABALLERO P, BANCHS A, DE VECIANA G, et al. Network Slicing Games: Enabling Customization in Multi⁃Tenant Networks [C]//2017 IEEE Conference on Computer Communications (INFOCOM). Atlanta, USA: 1–9. DOI: 10.1109/INFOCOM.2017.8057046. DOI
14	JIANG M, CONDOLUCI M, MAHMOODI T. Network Slicing in 5G: An Auction⁃Based Model[C]//2017 IEEE International Conference on Communications (ICC), Paris, France, 2017: 1–6. DOI: 10.1109/ICC.2017.7996490 DOI
15	WANG G, FENG G, Tan W, et al. Resource Allocation for Network Slices in 5G with Network Resource Pricing [C]//2017 IEEE Global Communications Conference (GLOBECOM). Singapore, Singapore, 2017: 1–6. DOI: 10.1109/GLOCOM.2017.8254074 DOI
16	MARQUEZ C, GRAMAGLIA M, FIORE M, et al. How Should I Slice My Network?: A Multi⁃Service Empirical Evaluation of Resource Sharing Efficiency [C]//24th Annual International Conference on Mobile Computing and Networking (MobiCom’18). New Delhi, India, 2018: 191–206. DOI: 10.1145/3241539.3241567 DOI
17	ZHENG J, CABALLERO P, DE VECIANA G, et al. Statistical Multiplexing and Traffic Shaping Games for Network Slicing [C]//15th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt). Paris, France, 2017: 1–8. DOI: 10.23919/WIOPT.2017.7959883 DOI
18	LECONTE M, PASCHOS G S, MERTIKOPOULOS P, et al. A Resource Allocation Framework for Network Slicing [C]//2018 IEEE Conference on Computer Communications (INFOCOM). Honolulu, USA, 2018: 2177–2185. DOI: 10.1109/INFOCOM.2018.8486303 DOI
19	CHANG C⁃Y, NIKAEIN N, SPYROPOULOS T. Radio Access Network Resource Slicing for Flexible Service Execution [C]//2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). Honolulu, USA, 2018: 668–673. DOI: 10.1109/INFCOMW.2018.8407021 DOI
20	HAN B, DE DOMENICO A, DANDACHI G, et al. Admission and Congestion Control for 5G Network Slicing [C]//2018 IEEE Conference on Standards for Communications and Networking (CSCN). Paris, France, 2018: 1–6. DOI: 10.1109/CSCN.2018.8581773 DOI
21	OLADEJO S O, FALOWO O E. Profit⁃Aware Resource Allocation for 5G Sliced Networks [C]//2018 European Conference on Networks and Communications (EuCNC). Ljublijana, Slovenia, 2018: 43–47. DOI: 10.1109/EuCNC.2018.8442661 DOI
22	HAN B, SCIANCALEPORE V, FENG D, et al. A Utility⁃Driven Multi⁃Queue Admission Control Solution for Network Slicing [C]//2019 IEEE Conference on Computer Communications (INFOCOM). Paris, France, 2019: 56–63. DOI: 10.1109/INFOCOM.2019.8737517 DOI
23	BEGA D, GRAMAGLIA M, FIORE M, et al. DeepCog: Cognitive Network Management in Sliced 5G Networks with Deep Learning [C]//2019 IEEE Conference on Computer Communications (INFOCOM). Paris, France, 2019: 280–288. DOI: 10.1109/INFOCOM.2019.8737488 DOI
24	SCIANCALEPORE V, DI RENZO M, COSTA⁃PERÉZ X. STORNS: Stochastic Radio Access Network Slicing [C]//2019 IEEE International Conference on Communications (ICC). Shanghai, China, 2019: 1–7
25	HAN B, FENG D, JI L, et al. A Profit⁃Maximizing Strategy of Network Resource Management for 5G Tenant Slices [EB/OL]. (2017-09-26).
26	HAN B, FENG D, SCIANCALEPORE V, et al. Rational Impatience Admission Control in 5G⁃Sliced Networks: Shall I Bide My Slice Opportunity? [EB/OL]. (2018-10-31).
27	CHATTERJEE K, W F (eds) SAMUELSON. Game Theory and Business Applications [M]. Boston, USA: Springer, 2002: 339–370
28	KRISHNA V. Auction Theory [M]. Cambridge, USA: Academic Press, 2009

Reference	Slice admission control	Cross-slice resource allocation	Policy-based	Auction-based	Note
[4]	N	Y	Y	N	User admission control on every individual slice according to tenant-specific policies to allocate resources cross slices.
[5]	N	Y	Y	Y	Policy-based user admission control and user dropping on every slice to guarantee QoS; auction-based intra-slice resource allocation among users; budget-based inter-slice resource allocation. Dynamic cross-slice resource allocation not considered.
[6]	N	Y	Y	N	Grouping users according to behaviors and social relationships; bio-inspired methods to update the groups; policy-based cross-slice resource allocation according to group status
[7]	Y	N	Y	N	Uniformed slice size, binary slice admission control according to the active slice set, genetic algorithm to optimize the policy.
[8]	N	Y	Y	N	Deep Q-learning assisted allocation policy optimization.
[9]	Y	N	Y	N	A Markov model for policy-based slice admission control.
[10]	N	Y	Y	N	Jointly optimizing the base station bandwidth and the backhaul capacity as a bi-convex problem.
[11]	N	Y	N	Y	Non-cooperative auction among slices for network resources, implemented with OpenFlow
[12]	Y	N	Y	N	Q-learning assisted slice admission control policy optimization.
[13]	N	Y	Y	Y	A preliminary conference version of[5].
[14]	N	Y	N	Y	A two-level slicing mechanism with 1) a price competition among network chunks to determine resource prices and 2) an auction mechanism to allocate resources among slices.
[15]	N	Y	N	Y	Optimizing the resource price function to maximize the total profit of slices / the net social welfare of network.
[16]	N	Y	Y	N	Empirical investigation on the diversity gain in SlaaS.
[17]	N	Y	Y	N	Sharing RAN resources among users according to both base station assignment and slice assignment. User admission control on every slice to shape traffic and guarantee the QoS.
[18]	N	Y	Y	N	Splitting the policy optimization problem into two sub-problems, one from the MNO’s perspective to maximize the revenue and the other on (every) tenant’s side to minimize the cost. A distributed optimization is therefore achieved through a game-fashion iteration of price updating. Both resource constraints and service fairness are taken account of.
[19]	N	Y	Y	N	Optimize the RAN resource allocation policy taking into account of the resource-partitioning problem.
[20]	Y	Y	Y	N	A two-layer framework merging slice admission control and cross-slice resource allocation.
[21]	N	Y	Y	N	Allocating users to subcarriers across different MVNOs to maximize the overall network profit, assuming the cost proportional to both power and bandwidth.
[22]	Y	N	Y	N	Multiple queues for different slice types, taking into account impatient behavior of tenants.
[23]	N	Y	Y	N	Dynamic resource allocation based on deep neural network assisted traffic prediction. Data-driven black-box optimization.
[24]	Y	Y	Y	N	Optimizing RAN resource allocation among slices and non-sliced network, where admissions to slice requests are controlled w.r.t. the demanded resource efficiency.
[25]	Y	N	Y	N	Modeling MNO’s revenue under policy-based slice admission control, analyzing the construction of optimal policy.
[26]	Y	N	Y	N	Studying the rational behavior of impatient tenants in policy-based slice admission control with multiple queues.

Reference	Slice admission control	Cross-slice resource allocation	Policy-based	Auction-based	Note
[4]	N	Y	Y	N	User admission control on every individual slice according to tenant-specific policies to allocate resources cross slices.
[5]	N	Y	Y	Y	Policy-based user admission control and user dropping on every slice to guarantee QoS; auction-based intra-slice resource allocation among users; budget-based inter-slice resource allocation. Dynamic cross-slice resource allocation not considered.
[6]	N	Y	Y	N	Grouping users according to behaviors and social relationships; bio-inspired methods to update the groups; policy-based cross-slice resource allocation according to group status
[7]	Y	N	Y	N	Uniformed slice size, binary slice admission control according to the active slice set, genetic algorithm to optimize the policy.
[8]	N	Y	Y	N	Deep Q-learning assisted allocation policy optimization.
[9]	Y	N	Y	N	A Markov model for policy-based slice admission control.
[10]	N	Y	Y	N	Jointly optimizing the base station bandwidth and the backhaul capacity as a bi-convex problem.
[11]	N	Y	N	Y	Non-cooperative auction among slices for network resources, implemented with OpenFlow
[12]	Y	N	Y	N	Q-learning assisted slice admission control policy optimization.
[13]	N	Y	Y	Y	A preliminary conference version of[5].
[14]	N	Y	N	Y	A two-level slicing mechanism with 1) a price competition among network chunks to determine resource prices and 2) an auction mechanism to allocate resources among slices.
[15]	N	Y	N	Y	Optimizing the resource price function to maximize the total profit of slices / the net social welfare of network.
[16]	N	Y	Y	N	Empirical investigation on the diversity gain in SlaaS.
[17]	N	Y	Y	N	Sharing RAN resources among users according to both base station assignment and slice assignment. User admission control on every slice to shape traffic and guarantee the QoS.
[18]	N	Y	Y	N	Splitting the policy optimization problem into two sub-problems, one from the MNO’s perspective to maximize the revenue and the other on (every) tenant’s side to minimize the cost. A distributed optimization is therefore achieved through a game-fashion iteration of price updating. Both resource constraints and service fairness are taken account of.
[19]	N	Y	Y	N	Optimize the RAN resource allocation policy taking into account of the resource-partitioning problem.
[20]	Y	Y	Y	N	A two-layer framework merging slice admission control and cross-slice resource allocation.
[21]	N	Y	Y	N	Allocating users to subcarriers across different MVNOs to maximize the overall network profit, assuming the cost proportional to both power and bandwidth.
[22]	Y	N	Y	N	Multiple queues for different slice types, taking into account impatient behavior of tenants.
[23]	N	Y	Y	N	Dynamic resource allocation based on deep neural network assisted traffic prediction. Data-driven black-box optimization.
[24]	Y	Y	Y	N	Optimizing RAN resource allocation among slices and non-sliced network, where admissions to slice requests are controlled w.r.t. the demanded resource efficiency.
[25]	Y	N	Y	N	Modeling MNO’s revenue under policy-based slice admission control, analyzing the construction of optimal policy.
[26]	Y	N	Y	N	Studying the rational behavior of impatient tenants in policy-based slice admission control with multiple queues.

[1]	LI Hanwen, BI Ningjing, SHA Jin. Design of Raptor-Like LDPC Codes and High Throughput Decoder Towards 100 Gbit/s Throughput [J]. ZTE Communications, 2023, 21(3): 86-92.
[2]	DING Jianwen, LIU Yao, LIAO Hongjian, SUN Bin, WANG Wei. Statistical Model of Path Loss for Railway 5G Marshalling Yard Scenario [J]. ZTE Communications, 2023, 21(3): 117-122.
[3]	CHEN Jiajun, GAO Yin, LIU Zhuang, LI Dapeng. Future Vision on Artificial Intelligence Assisted Green Energy Efficiency Network [J]. ZTE Communications, 2023, 21(2): 34-39.
[4]	AWADA Uchechukwu, ZHANG Jiankang, CHEN Sheng, LI Shuangzhi, YANG Shouyi. Machine Learning Driven Latency Optimization for Internet of Things Applications in Edge Computing [J]. ZTE Communications, 2023, 21(2): 40-52.
[5]	SHI Xiangyi, HAN Tongzhou, TIAN Hai, ZHAO Danfeng. Design of Raptor-Like Rate Compatible SC-LDPC Codes [J]. ZTE Communications, 2022, 20(S1): 16-21.
[6]	ZHAO Zipiao, ZHAO Yongli, YAN Boyuan, WANG Dajiang. Auxiliary Fault Location on Commercial Equipment Based on Supervised Machine Learning [J]. ZTE Communications, 2022, 20(S1): 7-15.
[7]	NAN Yucen, FANG Minghao, ZOU Xiaojing, DOU Yutao, Albert Y. ZOMAYA. A Collaborative Medical Diagnosis System Without Sharing Patient Data [J]. ZTE Communications, 2022, 20(3): 3-16.
[8]	ZHANG Jintao, HE Zhenqing, RUI Hua, XU Xiaojing. Spectrum Sensing for OFDMA Using Multicarrier Covariance Matrix Aware CNN [J]. ZTE Communications, 2022, 20(3): 61-69.
[9]	HOU Xiaolin, LI Xiang, WANG Xin, CHEN Lan, SUYAMA Satoshi. Some Observations and Thoughts about Reconfigurable Intelligent Surface Application for 5G Evolution and 6G [J]. ZTE Communications, 2022, 20(1): 14-20.
[10]	YAN Xincheng, TENG Huiyun, PING Li, JIANG Zhihong, ZHOU Na. Study on Security of 5G and Satellite Converged Communication Network [J]. ZTE Communications, 2021, 19(4): 79-89.
[11]	LIU Haipeng, ZHANG Xingyue, ZHOU Anfu, LIU Liang, MA Huadong. Indoor Environment and Human Sensing via Millimeter Wave Radio: A Review [J]. ZTE Communications, 2021, 19(3): 22-29.
[12]	XIAO Kai, LIU Xing, HAN Xianghui, HAO Peng, ZHANG Junfeng, ZHOU Dong, WEI Xingguang. Flexible Multiplexing Mechanism for Coexistence of URLLC and EMBB Services in 5G Networks [J]. ZTE Communications, 2021, 19(2): 82-90.
[13]	JI Hong, ZHANG Tianxiang, ZHANG Kai, WANG Wanyuan, WU Weiwei. Efficient Network Slicing with Dynamic Resource Allocation [J]. ZTE Communications, 2021, 19(1): 11-19.
[14]	LIU Zhuang, GAO Yin, LI Dapeng, CHEN Jiajun, HAN Jiren. Enabling Energy Efficiency in 5G Network [J]. ZTE Communications, 2021, 19(1): 20-29.
[15]	ZHANG Jing, WEI Xiao, CHENG Junfeng, FENG Xu. Satellite E2E Network Slicing Based on 5G Technology [J]. ZTE Communications, 2020, 18(4): 26-33.