1 / 67

Cross-Layer Network Planning and Performance Optimization Algorithms for Wireless Networks

Dissertation Proposal. Cross-Layer Network Planning and Performance Optimization Algorithms for Wireless Networks. Yean-Fu Wen Advisor: Frank Yeong-Sung Lin Department of Information Management, National Taiwan University 2007/4/24. Agenda. Introduction. Ch.2. Ch. 3.

ojal
Télécharger la présentation

Cross-Layer Network Planning and Performance Optimization Algorithms for Wireless Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Dissertation Proposal Cross-Layer Network Planning and Performance Optimization Algorithms for Wireless Networks Yean-Fu Wen Advisor: Frank Yeong-Sung Lin Department of Information Management, National Taiwan University 2007/4/24

  2. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast (or Multicast) • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work

  3. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Background • Wireless networks are the key to improving • person-to-person communications, • person-to-machine communications, and • machine-to-machine communications. • The research scope of this dissertation covers • various network architectures, and • various protocol layers [Ref: B3G Planning]

  4. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fairness model Wire Network (Fiber, T3…etc.) AP5 AP1 Network MAC layer AP2 AP6 PHY layer Application AP7 Network AP3 AP8 Mesh Networks MAC layer AP9 AP4 PHY layer MDE MDA MDD MAC layer MDB MDC PHY layer BS-oriented Ad Hoc, Sensor or Hybrid Networks

  5. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Motivation • Fairness • to ensure the allocated resources are sufficient for all MDs to achieve equivalent throughput or channel access time, and minimize end-to-end delay • to distribute and balance the traffic load or on related links • to solve fairness issues due to spatial bias or energy constraintsin three networks with different structures • Multi-range • causes different levels of energy consumption • causes different bit-rates (capacity) • Multi-rate • causes performance anomalies [Heusse’03]

  6. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Motivation • Multi-hop • causes throughput and end-to-end delay fairness issues in WMNs • causes inefficient energy usage in WSNs • Multicast • reduce the number of duplicate packets in order to gain a “multicast wireless advantage” and thereby reach multiple relay nodes • reduce the number of duplicate packets in data-centric WSNs • Multi-channel vs. Multi-access • whether to use multi-channel to reduce the number of collisions • Multi-sink • in WMNs, find a TAP trade-off in routing to a backhaul via a shorter path or routing to light-load links and a backhaul • in WSNs, find a source sensor trade-off between the shortest relay node or the sink node and the in-network process to reduce energy consumption

  7. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Objective • How to achieve throughput and channel access time fairness. • How to fairly allocate resources to solve the spatial bias problem in single hop or multi-hop wireless networks. • How to fairly distribute the traffic load among relay nodes to reduce end-to-end delay and among sensor nodes to increase the sensor network’s lifetime.

  8. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Research Approaches • Discrete event simulation [Harrell’92] • NS2 [Fall’99] • Analytic heuristic modeling [Harrell’92] • MATLAB • Lagrangean Relaxation (LR) [Ahuja’93] [Fisher’81] =0.5 =0.25 =0.125 … =1 =2

  9. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work Publication List [1] Y.F. Wen, Frank Y.S. Lin, and K.W. Lai,"System Throughput Maximization Subject to Delay and Time Fairness Constraints in 802.11 WLANs,"in Proc. of IEEEICPADS,Fukuoka Institute of Technology (FIT), Fukuoka, Japan, Jul. 2005. (EI) [2] Yu-Liang Kuo, Kun-Wai. Lai, Frank Yeong-Sung Lin, Yean-Fu Wen, Eric Hsiao-kuang Wu, and Gen-Huey Chen, "Multi-Rate Throughput Optimization with Fairness Constraints in Wireless Local Area Networks,"IEEE Transactions on Vehicular Technology, Dec. 2006 (major revised).

  10. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Ts Ts Tf Tf F Slow MD F Slow MD t Throughput fairness vs. channel access time fairness System Throughput Maximization Subject to Time Fairness Constraints in WLANs • We discuss how to achieve a trade-off between throughput fairness and channel access timefairness in 802.11 WLANs. • Problem • multiple bit-rates cause performance anomalies [Heusse’03]. Given:

  11. Objective: to maximize system throughput. Subject to: initial contention window size; packet size; multiple back-to-back packets; maximum cycle time (delay) time fairness To determine: the initial contention window size for each bit rate class, the packet size for each bit rate class, the number of multiple back-to-back packets of class-k Bkin a block within one transmission cycle Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion System Throughput Maximization Subject to Time Fairness Constraints in WLANs a great deal of computing time & non-convex problem T(N) DIFS SLOT SIFS data ACK t backoff time

  12. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion System Throughput Maximization Subject to Time Fairness Constraints in WLANs • Proposed algorithm • Modified binary search (Unimodal curve interval based on fairness index constraints[Jain’84] ) • Theorem: If the time value x is deducted from a class-k MH, and it does not change any other class-j MHs, then the fairness • increases iff x < xk – xj. • remains the same iff x = xk – xj. • decreases iff x > xk – xj.

  13. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion System Throughput Maximization Subject to Time Fairness Constraints in WLANs • Experiment results • although the problem has been shown to be NP-complete [Kuo’05], our numerical results reveal a simple unimodal feature • the relation between three MAC layer parameters (i.e., the initial contention window, packet size, and multiple back-to-back packets) and fairness achieves access time near-fairness and maximizes the system throughput with a simultaneous delay bound [Wen’06c]. • 21%improvement in system throughputover the original MAC protocol.

  14. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion System Throughput Maximization Subject to Time Fairness Constraints in WLANs • Related work • “performance anomaly” [Heusse’03] (Grenoble) • 802.11 system throughput analysis [Bianchi’00] • performance analysis under a finite load and improvements for multirate 802.11b [Cantieni’05] (Brunel) • to discuss the issues of cycle time (delay) [Wang’03], [Wu’02], [Chatzimisios’03], and [Raptis’05] • Jain’s Fairness Index (FI) model [Jain’84] • integer programming [Lai’04] [Kuo’05] (NTU) • an uplink solution with packet size or burst packets [Tan’04a] [Tan’04b] (MIT) • simulate a high quality signal with multiple back-to-back packets [Sadeghi’02] (Rice) [Sheu’02] (NCU)

  15. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work Publication List [1] Yean-Fu Wen and Frank Yeong-Sung Lin, "Fair Bandwidth Allocation and End-to-End Delay Routing Algorithms in Wireless Mesh Networks," Communications, IEICE Transactions on, E90-B(5), pp. xx–xx, May 2007. (SCI, EI)

  16. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Throughput and End-to-End Delay with Resource Allocation for WMNs • We discuss the scenario where many clients use the same backhaul to access the Internet. Consequently, throughput depends on each client’s distance from the gateway node. [Karrer’04] [Gambiroza’04] Given:

  17. Objective: to minimize the maximal end-to-end delay of the WMN. Subject to: capacity delay To determine: the resources cs(u,v) that should beallocated to the selected links of a TAP node s. the end-to-end delay on the selected path of a TAP node. the maximum end-to-end delayd of the WMN. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Throughput and End-to-End Delay with Resource Allocation for WMNs

  18. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion g s , 1 1 Link (1,3) 1 3 4 Link (3,4) Link (2,3) 2 g s , 2 2 Fair Throughput and End-to-End Delay with Resource Allocation for WMNs • Lemma 3-1: fair end-to-end delay is achievable • monotonic increases in f(u,v) • the delay time approaches ∞, when f(u,v) C(u,v) • the delay function is a convex function

  19. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Throughput and End-to-End Delay with Resource Allocation for WMNs • Experiment results

  20. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Throughput and End-to-End Delay with Resource Allocation for WMNs • Related work • wireless mesh networks: a survey [Akyildiz’05] (GIT) • describe 10 challenging issues [Karrer’04] (Rice university) • spatial bias fairness & temporal fairness [Gambiroza’04] (Rice university) • average delay, end-to-end delay routing and capacity assignment for virtual circuit networks [Cheng’95] [Yen’01] (NTU) • to maximize spatial reuse of a spectrum by maintaining basic fairness among contending flows [Li’05] (Toronto) • hierarchically aggregated fair queuing (HAFQ) for per-flow fair bandwidth allocation [Maki’06] (Osaka)

  21. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work Publication List [1] Frank Yeong-Sung Lin and Yean-Fu Wen, "Fair Inter-TAP Routing and Backhaul Assignment in Wireless Mesh Networks," was submitted toJournal of WCMC, Oct. 2006. (under review) [2] Y.F. Wen and Frank Y.S. Lin, "The Top Load Balancing Forest Routing in Mesh Networks,"in Proc. of IEEE CCNC, Las Vegas, NV, Jan. 2006. (EI)

  22. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs • How to cluster backbone mesh networks efficiently so that the load-balancedrouting is concentrated on given and “to-be-determined” backhauls. • Problem Given: backhaul TAP link

  23. Objective: to minimize the sum of the aggregated flows of selected links Subject to: budget  backhaul assignmentB backhaul selection routing Psb link L, p(u,v), Hs capacity Cuv load balancing ’, ’’ To determine: which TAPshould be selected to be a backhaulb which backhaul should be selected for each TAP to transmit its data zsb. The routing path from a TAP to a backhaul xp. whether a link should be selected for the routing path y(u,v). aggregated flow on top-level selected link f(u,b). aggregated flow on each backhaul b. a top-level load-balanced forest. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs

  24. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs (9) (1) (2) (10) BA+MCP= NP-complete (3) (11) (4) (12) (5) (14) (6) (7) (13) (8) (15)

  25. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs • Proposed algorithm • weighted backhaul assignment (WBA) algorithm • greedy load-balanced routing (GLBR) algorithm

  26. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs • Experiment results • the load-balanced routing and backhaul assignment experiment results demonstrate that the GLBR plus WBA algorithms with the LR-based approach achieve a gap of 30% and outperform other algorithms by at least 10%

  27. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair Inter-TAP Routing and Backhaul Assignment Algorithms for WMNs • Related work • traditional AP assignment focuses on coverage of the service area [Tutschku’99], [Unbehaun’03], [Mathar’00], and [Fortune’95] • cluster-head assignment methods, such as max-min d-hop cluster [Amis’00], LCA [Baker’81] • multi-constrained path problem (MCP) is an NP-complete problem [Wang’96] (London) • a single sink to balance the traffic load on the incoming link of an egress node • a general tree structure [Hsiao’01] (Harvard) • sensor networks [Dai’03] (Colorado)

  28. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work Publication List [1] Frank Yeong-Sung Lin and Yean-Fu Wen, "Fair Inter-TAP Routing and Backhaul Assignment in Wireless Mesh Networks," was submitted toJournal of WCMC, Oct. 2006. (under review) [2] Yean-Fu Wen and Frank Yeong-Sung Lin, "Fair Bandwidth Allocation and End-to-End Delay Routing Algorithms in Wireless Mesh Networks," Communications, IEICE Transactions on, E90-B(5), pp. xx–xx, May 2007. (SCI, EI)

  29. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • How to cluster backbone mesh networks efficiently so that the load-balanced routing and fair end-to-end delay are concentrated on given backhauls. • Objective: • to minimize the maximum end-to-end delay • Subject to: • routing Ps • link (tree or mesh) L,p(u,v) • resource allocation C(u,v) • delay (including end-to-end delay) • To determine: • The routing path from a TAP to a backhaul xp. • whether a link should be selected for the routing path y(u,v). • the resource that should beallocated to the selected links of a TAP node. cs(u,v) • the maximum end-to-end delayd of the WMN.

  30. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion (IP3-3) (IP3-2) (3-3.1) (3-2.1) (3-3.2) (3-2.2) (3-2.3) (3-3.3) (3-2.4) (3-3.4) (3-3.5) (3-2.5) (3-3.6) (3-2.6) (3-3.7) (3-2.7) (3-3.8) (3-2.8) Fair End-to-End Delay and Load-Balanced Routing objective function Tree structure Mesh structure subject to: Steiner tree & Knapsack Problem = NP-complete

  31. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Lagrangean Relaxation (LR3-2) • Lagrangean dual problem (D3-2) Tree structure (LR3-2) subject to: (3-2.1), (3-2.3), (3-2.4), (3-2.5), and (3-2.7). objective function (D3-2) subject to

  32. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Lagrangean Relaxation (LR3-3) • Lagrangean dual problem (D3-3) Mesh structure (LR3-3) subject to: (3-3.1), (3-3.3), (3-3.4), (3-3.5), and (3-3.7). objective function (D3-3) subject to

  33. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Sub-problem (SUB3-2.1) is related to decision variable xp. objective function (SUB3-2.1) subject to (3-2.1) Each sub-problem of OD-pair, xp, is a shortest path problem solved by considering the link weight (3-2.1) a top-level load-balanced problem

  34. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Sub-problem , related to decision variable y(u,v), ys(u,v), and cs(u,v). (SUB3-2.2) (SUB3-3.2) Mesh structure Tree structure objective function (SUB3-2.2) (SUB3-3.2) subject to (3-2.3), (3-2.4), (3-2.5), and (3-2.7). subject to (3-3.3), (3-3.4), (3-3.5), and (3-3.7). we consider the delay function is M/M/1:

  35. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Sub-problem , related to decision variable d. (SUB3-3.3) (SUB3-2.3) objective function Mesh structure Tree structure (SUB3-2.3) (SUB3-3.3) subject to the lower and upper bound of d. Lemma 3-3: how to determine the upper bound and lower bound of d? As described in LR approach, the getting primal feasible solution gets the upper bound of this problem. The lower bound is calculated by the fully distributed the traffic load, aggregated from the higher level around the level of each node. (BFS)

  36. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Fair End-to-End Delay and Load-Balanced Routing • Getting primal feasible solution • the incoming link costs of the backhaul are set to delay function (u,v) = sSD(u,v)(C(u,v),s) * (a +), where (u,v) L. • greedy load-balanced routing (GLBR) [Wen’06a] • resource allocation scheme (EDTB) [Wen’07]

  37. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment Algorithms • Fair End-to-End Delay and Load-balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work Publication List [1] Frank Yeong-Sung Lin and Yean-Fu Wen, "A Path-Based Minimum Power Broadcast Algorithm in Wireless Networks," was submitted to ACM Baltzer Mobile Networks and Applications (MONET), Mar. 2007. (under review) [2] Frank Y.S. Lin, Y. F. Wen, L.C. Fu, and S.P. Lin, “A Path-Based Minimum Power Broadcast Problem in Wireless Networks," in Proc. of IEEE TENCON, Melbourne, Australia, Nov. 2005. (EI)

  38. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion A Minimum Power Broadcast Algorithm for Ad-hoc (Sensor) Networks • We discuss how to construct a multicast tree that minimizes power consumption with the “multicast wireless advantage”. • Problem Given: 8 ev(rv) 5 9 ¶ ( 1 , 2 ) 2 ¶ Power 10 1 ev(rv)=rv + a ( 1 , 3 ) consumption 6 3 (normalized) 11 rv 7 Power range 4 12

  39. Objective: to minimize the total broadcast power consumption Subject to: routing tree radius To determine: the routing path from each source to the destination, denoted as an OD-pair xp. whether a link should be on the multicast tree y(u,v) . a multicast tree. transmission radius for each MD ru. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion A Minimum Power Broadcast Algorithm for Ad-hoc (Sensor) Networks a multicast tree which is also a Steiner tree = NP-complete

  40. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion A Minimum Power Broadcast Algorithm for Ad-hoc (Sensor) Networks • Proposed algorithm • a minimum power broadcast algorithm • Experiment results

  41. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion A Minimum Power Broadcast Algorithm for Ad-hoc (Sensor) Networks • Related work • power range and topology control [Salhieh’01] [Bettstetter’02] [Santi’01] • to build minimum energy networks via the shortest path tree (SPT) algorithm, measuring the cost of the edge by its power level [Salhieh’01], [Dowell’01], [Montemanni’04], and [Li’01] • node-based solutions in static all-wireless networks in terms of a trade-off: a node can reach more nodes in a single hop by using higher transmission power [Ahluwalia’05], [Wieselthier’00] and [Cagalj’02] • link-based solutions [Das’03a] • a series of heuristics (e.g., BIP) to solve this problem [Wieselthier’00], [Wieselthier’01], and [Wieselthier’02] • r-shrink [Das’03b]

  42. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda Publication List [1] Frank Yeong-Sung Lin and Yean-Fu Wen, "Multi-sink Data Aggregation Routing and Scheduling with Dynamic Radii in WSNs," IEEE Communications Letters, 10(10), pp. 692–694, Oct. 2006. (SCI, EI) [2] Yean-Fu Wen and Frank Yeong-Sung. Lin, "Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing in Wireless Sensor Networks," Lecture Notes Computer Science (LNCS), vol. 4096, pp. 894–903. (Proceedings of IFIP EUC 2006) (SCI, EI) [3]Y.F. Wen, Frank Y.S. Lin, and W.C. Kuo, "A Tree-based Energy-efficient Algorithm for Data-Centric Wireless Sensor Networks,"in Proc. of IEEE AINA, Ontario, Canada, May 2007. (EI) [4]Y.F. Wen and Frank Y.S. Lin, "Energy-Efficient Data Aggregation Routing and Duty-Cycle Scheduling in Cluster-based Sensor Networks," in Proc. of IEEE CCNC, Las Vegas, NV, Jan. 2007. (EI) • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work

  43. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing for WSNs • We discuss how to increase the battery lifetime and energy consumption efficiency of a network from the Physical layer to the Application layer in terms of the following issues: • single/multi-sink • data aggregation • tree structure routing • duty-cycle scheduling • node-to-node communication time • the number of retransmissions • dynamically adjusted radius Application layer Network layer MAC layer Physical layer

  44. Objective: minimize the total energy consumed by a target transmission to one of sink nodes Subject to: sink selection D restrictions on the structure of trees in the form of three link constraints PsD,L, p(u,v),Hs duty cycle scheduling the time for node-to-node communication [Shiou’05] dynamic radius uv, Ru To determine: The sink node that a source node bsg will route to; a routing pathxp and linky(u,v) from the source node to the sink node; the earliest time nu at which a node wakes up and begins aggregating data; and the timemu at which aggregation of sub-tree data will be completed; the time uv needed for a successful node-to-node transmission. the power rangeru of each node; Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing for WSNs a type of reverse-multicast tree which is also a Steiner tree MCP = NP-complete

  45. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing for WSNs • Proposed algorithm: single sink 0 0 κ [3, 5+0] D 1 0 2 0 2 [3, 4+1] 1 0 [0, 3+2] 4 [0, 3+1] 3 1 3 2 [0, 0+1] 5 [0, 0+3] [0, 0+2] 6 7 S1 3 S2 [0, 0+3] S3 8 ∞ 0 ∞ 0 S4 0 ∞ ∞ 0 O

  46. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Energy-Efficient Data Aggregation Routing and Duty-Cycle Scheduling for Multi-Sink WSNs • Proposed algorithm: multi-sink • We discuss how to increase the lifetime of the networks already discussed with a multiple sink structure (outgoing information gateways)

  47. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing for WSNs • Experiment results: single sink

  48. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Energy-Efficient Data Aggregation Routing and Duty-Cycle Scheduling for Multi-Sink WSNs • Experiment results: multi-sink

  49. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Cross-Layer Duty Cycle Scheduling with Data Aggregation Routing for WSNs • Related work • backhaul selection [Wen’07] (NTU) • multi-sink [Yuen’06] (Toronto), [Kalantari’06] (Maryland), [Kim’06] (Seoul) • three aggregation heuristics, namely, the Shortest Paths Tree (SPT), Center at Nearest Source (CNS), and the Greedy Incremental Tree (GIT) [Krishnamachari’02] (USC) • the tradeoff between power consumption and coverage of transmission nodes [Carle’04] • S-MAC [Ye’02], T-MAC [Dam’03], D-MAC [Lu’04a] [Lu’04b] • retransmission [Shiou’05] [Bianchi’00] [Sheu’03] [Wen’06c] • radius (refer to Ch. 5)

  50. Agenda Introduction Ch.2 Ch. 3 Ch. 4 Ch.3 Ch.5 Ch.6,7 Conclusion Agenda • Introduction • Wi-Fi Hotspots(Ch. 2) • System Throughput Maximization Subject to Time Fairness Constraints • Wireless Mesh Networks(Ch. 3 and Ch. 4) • Fair Throughput and End-to-End Delay with Resource Allocation • Fair Inter-TAP Routing and Backhaul Assignment • Fair End-to-End Delay and Load-Balanced Routing • Ad Hoc Networks(Ch. 5) • A Minimum Power Broadcast Algorithm • Wireless Sensor Networks(Ch. 6 and Ch. 7) • Dynamic Radius, Duty Cycle Scheduling, Routing, Data Aggregation, and Multi-Sink (Clusters) • Conclusion & Future Work

More Related