考慮端對端延遲與流量公平性之無線網狀網路最佳化建置 Deployment Optimization of Wireless Mesh Networks Considering End-to-End Delay and Throughput Fairness 考慮端對端延遲與流量公平性之無線網狀網路最佳化建置 蔡永斌 Tsai, Yung-Pin 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Agenda Introduction Problem Formulation Background Motivation Literature Survey Problem Formulation Problem Description 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Background Taxonomy of Wireless Networks Taxonomy of Wireless Mesh Networks 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Taxonomy of Wireless Networks Networking Multi-hop Infrastructure-less Single-hop Wireless Wide Area Networks (WWANs) Mesh Networks (WMNs) Infrastructure-based Wireless Local (WLANs) Metropolitan (WMANs) Personal Area Networks (WPANs) 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Infrastructure-based Taxonomy of WMNs WMNs Infrastructure-based Infrastructure-less Hybrid Infrastructure/ backbone WMNs Client WMNs Hybrid WMNs 分類方法參考自: I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Infrastructure/backbone WMNs 圖擷取自: I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Client WMNs 圖擷取自: I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Hybrid WMNs 圖擷取自: I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Agenda Introduction Problem Formulation Background Motivation Literature Survey Problem Formulation Problem Description 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Motivation Performance limitation of multi-hop wireless networks: Channel bandwidth is shared by a number of transmissions. Benefit gained through building relay nodes decreases when network grows. [1] Longer hop paths must bear significantly lower throughput compare with short paths. [2][3][4][5] The performance of the network can improve is limited, because the total network capacity has been constrained by network architecture. [1] P. Gupta and P. R. Kumar, “The Capacity of Wireless Networks,” IEEE Transactions on information theory, Vol. 46, No. 2, Pages 388-404, 2000. [2] I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. [3] B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. [4] V. Gambiroza, B. Sadeghi and E. W. Knightly, “End-to-End Performance and Fairness in Multihop Wireless Backhaul Networks,” Proc. ACM MobiCom, Pages 287-301, Sept.-Oct. 2004. [5] Y. F. Wen and Frank Y. S. Lin, “The Top Load Balancing Forest Routing in Mesh Networks”, Proc. IEEE CCNC, pages 468-472, Jan. 2006. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Motivation Even through during the operation stage we can deploy additional equipments or upgrade equipments to increase link capacity, throughput is not always raised[1]. The greatest motivation of our work: We consider end-to-end throughput and delay fairness at network planning stage simultaneously. Optimization issues like channel assignment, routing and flow control are jointly considered in the WMNs deployment problem we addressed. With the objective to maximize the minimal end-to-end throughput, in other word, to maximize the total network capacity. [1] A. Tang, J. Wang and S. H. Low, “Counter-Intuitive Throughput Behaviors ini Networks Under End-to-End Control,” IEEE Transactions on Networking, Vol. 14, No. 2, Pages 355-368, April 2006. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Agenda Introduction Problem Formulation Background Motivation Literature Survey Problem Formulation Problem Description 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Literature Survey Capacity of WMNs End-to-End Fairness of WMNs IEEE 802.11s/D3 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Capacity of WMNs The capacity of WMNs is influenced by many factors including : Network architecture Network topology Traffic pattern Network node density Number of channels used for each node Transmission power level I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Capacity of WMNs Capacity analysis of multi-hop wireless networks [1] P. Gupta and P. R. Kumar Optimal placed and random placed multi-hop wireless networks Interference models: physical model and protocol model Transmission power control Tradeoff between spatial reuse efficiency and end-to-end delay. [2] When a node has six neighboring nodes, there will be optimum transmission power level. [3,4] Flexible transmission power control capability can significantly increase network performance. [1] I. F. Akyildiz, X. Wang and W. Wang, “Wireless Mesh Networks: a Survey,” Computer Networks, Vol. 47, Issue 4, Pages 445-487, March 2005. [2] L. Kleinrock and J. Silvester, “Optimum Transmission Radii for Packet Radio Networks or Why Six Is A Magic Number,” Proc. IEEE National Telecommunications Conference, Pages 4.3.1-4.2.5, December 1978. [3] E. M. Royer, P. M. M. Smith and L. E. Moser, “An Analysis of the Optimum Node Density for Ad hoc Mobile Networks,” Proc. IEEE ICC, Vol. 3, Pages 857-861, June 2001. [4] J. Gomez and A. T. Campbell, “Variable-Range Transmission Power Control in Wireless Ad Hoc Networks,” IEEE Transaction on Mobile Computing, Vol. 6, No. 1, Pages 87-99, January 2007. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Capacity of WMNs To increase the network capacity: End-to-end fairness Multi-hop multi-channel multi-radio wireless networks [1, 2, 3, 4, 5] Joint channel assignment, link scheduling, routing algorithms [1, 2, 4, 5] End-to-end fairness The tradeoff between fair quality of service (QoS) and network performance [6] [1] H. Yu, P. Mohapatra and X. Lin, “Channel Assignment and Link Scheduling in Multi-Radio Multi-Channel Wireless Mesh Networks,” Mobile Networks and Applications, Vol. 13, Issue 1-2, Pages 169-185, April 2008. [2] M. Kodialam and T. Nandagopal, “Characterizing the Capacity Region in Multi-Radio Multi-Channel Wireless Mesh Networks,” Proc. MobiCom’05. [3] W. Wang and X. Liu, “A Framework for Maximum Capacity in Multi-channel Multi-radio Wireless Networks,” Proc. IEEE CCNC’06. [4] X. Y. Li, A. Nusairat, Y. Wu, Y. Qi, J. Zhao, X. Chu and Y. Liu, “Joint Throughput Optimization for Wireless Mesh Networks,” IEEE Transactions on Mobile Computing, Vol. 8, Issue 7, Pages 895-909, July 2009. [5] J. Tang, G. L. Xue and W. Y. Zhang, “Cross-Layer Optimization For End-To-End Rate Allocation in Multi-Radio Wireless Mesh Networks,” Wireless Networks, Vol. 15, No. 1, January 2009. [6] B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
End-to-End Fairness of WMNs In WMNs, competitions for resources include inter-flow contention and intra-flow contention [1]. Max-min fairness model The fairness issue studied in our work includes: End-to-end throughput fairness End-to-end delay fairness [1] B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Inter-Flow Contention In a single-hop wireless network: w2 : w1 = 1 : 2 F1 F2 2/3B 1/3B 參考自: B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Intra-Flow Contention In a multi-hop wireless mesh network F1 F2.1 F2.2 F2.3 2/3B 1/9B r2 : r1 = 1 : 6 unfair 1/9B 1/3B 1/9B 參考自: B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. 2/5B 1/5B r2 : r1 = 1 : 2 FAIR 1/5B 3/5B 1/5B 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
End-to-End Fairness of WMNs To achieve end-to-end throughput fairness: Bandwidth allocation [1] Spatial reuse maximization The method extends the per-hop link bandwidth allocation to multi-hop flow fairness. This kind of approaches can be achieved through flow control To achieve end-to-end delay fairness: Load-balanced routing [2] End-to-end aggregated mean delay and delay jitter have been used as the routing metric Bandwidth allocation and load-balanced routing [2] [1] B. Li, “End-to-End Fair Bandwidth Allocation in Multi-Hop Wireless Ad Hoc Networks,” Proc. IEEE ICDCS’05. [2] Y.F. Wen, F.Y.S. Lin, Y.C. Tzeng and C.T. Lee, "Backhaul Assignment and Routing Algorithms with End-to-End QoS Constraints for Wireless Mesh Networks", Springer Wireless Personal Communications, Feb. 2009. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
End-to-End Fairness of WMNs Cross-layer optimization approach [1] Rate allocation, routing, scheduling, transmission power control and channel assignment problems Max-min fairness (end-to-end throughput fairness) Multi-hop wireless network deployment [2] Backhaul and routing assignment problem which guarantees end-to-end delay fairness. In our research: Mesh router and backhaul assignment, power control, channel assignment, routing and flow control Both end-to-end throughput and end-to-end delay fairness [1] J. Tang, G. L. Xue and W. Y. Zhang, “Cross-Layer Optimization For End-To-End Rate Allocation in Multi-Radio Wireless Mesh Networks,” Wireless Networks, Vol. 15, No. 1, January 2009. [2] Y.F. Wen, F.Y.S. Lin, Y.C. Tzeng and C.T. Lee, "Backhaul Assignment and Routing Algorithms with End-to-End QoS Constraints for Wireless Mesh Networks", Springer Wireless Personal Communications, Feb. 2009. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
IEEE 802.11s IEEE 802.11s/D3 Internet An enhancement of the original standard to support WLAN implementations for more flexible interoperable wireless connectivity. Internet Portal 1 Portal 2 AP MC1 MC2 MC3 MC4 MC5 MBSS infrastructure BSS Mesh STA 3 Mesh STA 7 Mesh STA 2 Mesh STA 6 Mesh STA 1 Mesh STA 8 Mesh STA 5 Mesh STA 4 圖文參考自: IEEE Standard, “Draft STANDARD for Information Technology- Telecommunications and information exchange- between systems- Local and metropolitan area networks- Specific requirements- Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications,” March 2009. infrastructure BSS 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Agenda Introduction Problem Formulation Background Motivation Literature Survey Problem Formulation Problem Description 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Description Objective in this paper is to establish an infrastructure WMN and make the network total network capacity as large as we can with end-to-end fairness requirements. A dominated node represents Internet Mesh STAs Portals 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
System Model Decisions Given Dominated node Mesh STAs with portal functionalities Decisions Mesh STAs with AP functionalities Given Mobile clients 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Protocol Interference Model Suppose node transmits to a node . Then this transmission is successfully received by node if for every other node simultaneously transmitting over the same channel k 參考自: P. Gupta and P. R. Kumar, “The Capacity of Wireless Networks,” IEEE Transactions on information theory, Vol. 46, No. 2, Pages 388-404, 2000. i j 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Contentions Between Flows F1 and F2 are two different end-to-end flows. Dash lines represented as intra-flow contention happens between F1 and F2 Inter-Flow Contention F1 F2.1 F2.2 F2.3 F1 and F2 are two different end-to-end flows. Intra-flow contentions between sub-flows F2.1, F2.2 and F2.3 Intra-Flow Contention 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Description Routing Uplink Unicasting Single path Dominated node Mesh STAs with portal functionalities Mesh STAs with AP functionalities Mobile clients 2018/12/4
Problem Formulation Assumptions Each mesh STA is stationary. Each mesh STA can be equipped with multiple radios each of which operates on a particular and non-overlapping channel. Radios of each mesh STA are omni-directional. The sources are greedy and the flow from each source is backlogged. Packets can be buffered at mesh STAs while awaiting transmission. The routing path of each OD-pair is single path. Each mesh STA has AP functionalities and can be upgrade to have portal functionalities. Each MC must home to a mesh STA and the selection is through system. Each MC and mesh STA has the capability of perfect power control. The channel MCs use is selected by system. 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Given CLs for mesh STAs Weighting factors for each MC Cost function for building a mesh STA and additional cost to have portal functionalities Distances between nodes Maximum transmission range of each radio Limited number of available channels of each radio Budget constraint 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Objective: Subject to: To maximize the minimal weighted end-to-end throughput Subject to: Budget constraint Mesh STA and portal assignment constraints Routing constraints Link assignment constraints Link constraints Capacity constraints QoS constraints Integer constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Given Parameters 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Decision Variables 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Objective function: Budget constraint: LP1 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Mesh and portal assignment constraints: Routing constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Link Assignment Constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Link constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Capacity constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation QoS constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Problem Formulation Integer constraints 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University
Thanks for your listening 2018/12/4 Network Optimization Laboratory, Information Management Department, National Taiwan University