Nov. 15 2002Chung-Hsien Hsu1 Round Robin with Look Ahead: A New Scheduling Algorithm for Bluetooth Daqing Yang, Gouri Nair, Balaji Sivaramakrishnan, Harishkumar.

Slides:



Advertisements
Similar presentations
Ch. 12 Routing in Switched Networks
Advertisements

Online Algorithm Huaping Wang Apr.21
Doc.: IEEE /080r1 Submission January 2001 Jie Liang, Texas InstrumentsSlide 1 Jie Liang Texas Instruments Incorporated TI Blvd. Dallas,
Ch. 12 Routing in Switched Networks Routing in Packet Switched Networks Routing Algorithm Requirements –Correctness –Simplicity –Robustness--the.
CH 5. CPU Scheduling Basic Concepts F CPU Scheduling  context switching u CPU switching for another process u saving old PCB and loading.
A Centralized Scheduling Algorithm based on Multi-path Routing in WiMax Mesh Network Yang Cao, Zhimin Liu and Yi Yang International Conference on Wireless.
Min Song 1, Yanxiao Zhao 1, Jun Wang 1, E. K. Park 2 1 Old Dominion University, USA 2 University of Missouri at Kansas City, USA IEEE ICC 2009 A High Throughput.
Multicast in Wireless Mesh Network Xuan (William) Zhang Xun Shi.
1 A Flexible Bandwidth Management Scheme in Bluetooth Chin-Fu Liu Multimedia and Communications Laboratory Department of Computer Science and Information.
Sep. 18, 2003Chung-Hsien Hsu A Dynamic Slot Assignment Scheme for Slave-to-Slave and Multicast-like Communication in Bluetooth Personal Area Networks C.
Chapter 3: CPU Scheduling
Hasan SÖZER1 Data Scheduling and SAR for Bluetooth MAC Manish Kalia, Deepak Bansal, Rajeev Shorey IBM India Research Laboratory.
1 Token Bucket Based CAC and Packet Scheduling for IEEE Broadband Wireless Access Networks Chi-Hung Chiang
Energy-Efficient Rate Scheduling in Wireless Links A Geometric Approach Yashar Ganjali High Performance Networking Group Stanford University
Bluenet a New Scatternet Formation Scheme * Huseyin Ozgur Tan * Zifang Wang,Robert J.Thomas, Zygmunt Haas ECE Cornell Univ*
And now … Graphs simulation input file parameters 10,000 requests 4 categories of file sizes 1K- 80% frequency 4K – 15% 16K – 4% 64K –1% poisson arrival.
1 Queue Scheduling Analysis The Computer Communication Lab (236340) - Winter 2004 Tanya Berezner Ana Gluzband Gitit Amihud.
Scatternet Formation in Bluetooth CSC 457 Bill Scherer November 8, 2001.
Presented by Hasan SOZER1 A Pseudo Random Coordinated Scheduling Algorithm for Bluetooth Scatternets Andras Racz, Gyorgy Miklos, Ferenc Kubinszky, Andras.
Analysis of Input Queueing More complex system to analyze than output queueing case. In order to analyze it, we make a simplifying assumption of "heavy.
Pipelined Two Step Iterative Matching Algorithms for CIOQ Crossbar Switches Deng Pan and Yuanyuan Yang State University of New York, Stony Brook.
Localized Asynchronous Packet Scheduling for Buffered Crossbar Switches Deng Pan and Yuanyuan Yang State University of New York Stony Brook.
© 2010 AT&T Intellectual Property. All rights reserved. AT&T and the AT&T logo are trademarks of AT&T Intellectual Property. Bin Packing: From Theory to.
Modeling & Simulation of Bluetooth MAC protocol COE543 Term Project Spring 2003 Submitted by: H.M.Asif (ID# )
Simulation of Memory Management Using Paging Mechanism in Operating Systems Tarek M. Sobh and Yanchun Liu Presented by: Bei Wang University of Bridgeport.
CPU-Scheduling Whenever the CPU becomes idle, the operating system must select one of the processes in the ready queue to be executed. The short term scheduler.
Chapter 6: CPU Scheduling
OPERATING SYSTEMS CPU SCHEDULING.  Introduction to CPU scheduling Introduction to CPU scheduling  Dispatcher Dispatcher  Terms used in CPU scheduling.
OPTIMAL SERVER PROVISIONING AND FREQUENCY ADJUSTMENT IN SERVER CLUSTERS Presented by: Xinying Zheng 09/13/ XINYING ZHENG, YU CAI MICHIGAN TECHNOLOGICAL.
Providing Delay Guarantees in Bluetooth Rachid Ait Yaiz and Geert Heijenk International Conference on Distributed Computing Systems Workshops (ICDCSW’03)
© 2009 IBM Corporation 1 Improving Consolidation of Virtual Machines with Risk-aware Bandwidth Oversubscription in Compute Clouds Amir Epstein Joint work.
© 2010 AT&T Intellectual Property. All rights reserved. AT&T and the AT&T logo are trademarks of AT&T Intellectual Property. Case Studies: Bin Packing.
Energy-Optimal Online Algorithms for Broadcasting in Wireless Network Shay Kutten Hirotaka ono David Peleg Kunihiko Sadakane Masafumi Yamashita.
Energy-Saving Scheduling in IEEE e Networks Chia-Yen Lin, and Hsi-Lu Chao Department of Computer Science National Chiao Tung University.
Competitive Queue Policies for Differentiated Services Seminar in Packet Networks1 Competitive Queue Policies for Differentiated Services William.
BLUETOOTH WIRELESS TECHNOLOGY
STUMP: Exploiting Position Diversity in the Staggered TDMA Underwater MAC Protocol Kurtis Kredo II, Petar Djukic, Prasant Mohapatra IEEE INFOCOM 2009.
Parametric Optimization Of Some Critical Operating System Functions An Alternative Approach To The Study Of Operating Systems Design.
A Fast Handoff Scheme For IP over Bluetooth Sang-Hsu Chung, Hyunsoo Yoon, and Jung-Wan Cho Department of Electrical Engineering & Computer Science Divison.
Multi-Power-Level Energy Saving Management for Passive Optical Networks Speaker: Chia-Chih Chien Advisor: Dr. Ho-Ting Wu Date: 2015/03/25 1.
Token-DCF, COMSNET(2013) -> MOBICOM(2014). Introduction ▣ To improve standard MAC protocol of IEEE for WLAN. ▣ S-MAC, A-MAC, SPEED-MAC, and etc.
Doc.: IEEE /66r0 Submission March 2000 Nada Golmie, NISTSlide 1 IEEE P Working Group for Wireless Personal Area Networks WPAN Coexistence.
Path Capacity Estimation in Time-Slotted Wireless Networks
Throughput-Oriented MAC for Mobile Ad Hoc Networks with Variable Packet Sizes Fan Wang, Ossama Younis, and Marwan Krunz Department of Electrical & Computer.
Seungnam Yang, Kyungsoo Lee, Hyundoc Seo and Hyogon Kim Korea University VTC Spring 2008 SPLASH: a Simple Multi-Channel Migration Scheme for IEEE
A Pseudo Random Coordinated Scheduling Algorithm for Bluetooth Scatternets MobiHoc 2001.
1 A Cross-Layer Scheduling Algorithm With QoS Support in Wireless Networks Qingwen Liu, Student Member, IEEE, Xin Wang, Member, IEEE, and Georgios B. Giannakis,
1 Grid-Based Access Scheduling for Mobile Data Intensive Sensor Networks C.-K. Lin, V. Zadorozhny and P. Krishnamurthy IEEE International Conference on.
Fair and Efficient multihop Scheduling Algorithm for IEEE BWA Systems Daehyon Kim and Aura Ganz International Conference on Broadband Networks 2005.
2003/03/14 An Efficient Link Polling Policy by Pattern Matching for Bluetooth Piconets Ting-Yu Lin, Yu-Chee Tseng, and Yuan-Ting Lu Department of Computer.
SERENA: SchEduling RoutEr Nodes Activity in wireless ad hoc and sensor networks Pascale Minet and Saoucene Mahfoudh INRIA, Rocquencourt Le Chesnay.
A Two-Tier Heterogeneous Mobile Ad Hoc Network Architecture and Its Load-Balance Routing Problem C.-F. Huang, H.-W. Lee, and Y.-C. Tseng Department of.
University of Padova Department of Information Engineering On the Optimal Topology of Bluetooth Piconets: Roles Swapping Algorithms Daniele Miorandi &
A Flexible Scatternet- wide scheduling Algorithm for Bluetooth Wensheng Zhang and Guohong Cao IPCCC 2002.
CPU Scheduling Operating Systems CS 550. Last Time Deadlock Detection and Recovery Methods to handle deadlock – Ignore it! – Detect and Recover – Avoidance.
Scheduling Mechanisms Applied to Packets in a Network Flow CSC /15/03 By Chris Hare, Ricky Johnson, and Fulviu Borcan.
Agenda Background Project goals Project description –General –Implementation –Algorithms Simulation results –Charts –Conclusions.
CPU Scheduling Algorithms
Bin Packing Optimization
IEEE Student Paper Contest
Chapter 6: CPU Scheduling
Scheduling in Packet Networks
Scheduling Algorithms in Broad-Band Wireless Networks
ns-2 simulation of TCP + CBR traffic
Throughput-Optimal Broadcast in Dynamic Wireless Networks
CPU SCHEDULING.
Qingwen Liu, Student Member, IEEE Xin Wang, Member, IEEE,
Data Structures for Shaping and Scheduling
Switch Performance Analysis and Design Improvements
Presentation transcript:

Nov Chung-Hsien Hsu1 Round Robin with Look Ahead: A New Scheduling Algorithm for Bluetooth Daqing Yang, Gouri Nair, Balaji Sivaramakrishnan, Harishkumar Jayakumar and Arunabha Sen ICPPW’ IEEE Speaker: Chung-Hsien Hsu

Nov Chung-Hsien Hsu2 Outline  Introduction  Bin Packing Problem  Scheduling Algorithm –Look Ahead –Look Ahead Round Robin  Simulation  Conclusion

Nov Chung-Hsien Hsu3 Introduction SCO: Synchronous Connection Oriented ACL: Asynchronous Connectionless Links frame

Nov Chung-Hsien Hsu4 Introduction  Objective: –To achieve high channel utilization. (throughput).

Nov Chung-Hsien Hsu5 Bin Packing Problem  Bin Packing Problem: –A set of objects L = { a 1,…,a n } with a size s(a i ) –A set of bins with a fixed capacity B –Objective: To pack the objects a i, 1<=i<=n, into as few bins as possible.

Nov Chung-Hsien Hsu6 Bin Packing Problem (cont.)  Off-line bin packing problem: –Have the complete knowledge of all elements and their sizes. Efficiently pack the objects in the bins.  Online bin packing problem: –Limited amount of future knowledge. –The scheduling algorithm for Bluetooth

Nov Chung-Hsien Hsu7 Bin Packing Problem (cont.) M S1S1 S2S2 MQ 1 ( 5,5,5,5,5)MQ 2 ( 1,1,1,1,1) SQ 1 ( 5,5,5,5,5) SQ 2 ( 1,1,1,1,1) CMSQ 1 ( 10,10,10,10,10)CMSQ 2 ( 2,2,2,2,2) Bluetooth:

Nov Chung-Hsien Hsu8 Bin Packing Problem (cont.)  Supposition: –The number of slots between every SCO slot pair is 10. Frame size = 10 As a bin with capacity 10.  The packets in the CMSQ can be viewed a objects.

Nov Chung-Hsien Hsu9 Scheduling Algorithm  Definition: –Asymptotic worst-case performance ratio The minimum number of bins required to pack list L The number of bins used by packing list L Scheduling Algorithm A L = (a 1,a 2,…,a n ) : the packets being scheduled by algorithm A

Nov Chung-Hsien Hsu10 Scheduling Algorithm (cont.)  Theorem: –For the Round Robin scheduling algorithm A in Bluetooth, R A = 5/3 Proof: Bin sizes: 10 Item sizes: 2, 4, 6, 8, 10. The Round Robin algorithm A guaranteed that any neighboring pair of bins packed item sizes at least 12. If OPT(L) = n  the packets need to be packed have size at most 10n. A(L) <= 20n/12 R A = 5/3 The Round Robin algorithm A compared with any scheduling algorithm B in Bluetooth, A is at most 5/3 worse than B.

Nov Chung-Hsien Hsu11 Scheduling Algorithm (cont.)  Look Ahead  Look Ahead Round Robin

Nov Chung-Hsien Hsu12 Look Ahead Step 1: Examine the (Head-Of-the-Line) packets CMSQ 1,…CMSQ x for all the active slaves in the piconet. Step 2: Prioritize slaves in order of non-increasing size of their HOL packets in CMSQ (the slave with the largest HOL packet has the highest priority). Step 3: If possible, schedule the highest priority slave in the current frame. Step 4: If HOL packet of the highest priority does not fit, search through HOL priority list for highest priority that will fit the current frame. Step 5: If none fits, wait for the start of a new frame and put the HOL packet in the new frame. Repeat step 1 through 5.

Nov Chung-Hsien Hsu13 Look Ahead (cont.)  Drawback –Computational overhead may be unacceptably high. If N slaves, it can be computed in O(log N) amount of computation. –Starvation.

Nov Chung-Hsien Hsu14 Look Ahead Round Robin  Round Robin + Look Ahead. –To avoid the starvation. –To reduce the computational complexity.  Different from RR: –When the current packet does not fit, the algorithm looks ahead and attempt to schedule a packet from next slave in line.

Nov Chung-Hsien Hsu15 Look Ahead Round Robin (cont.) Step 1: The algorithm will fill the current frame with the HOL packets in a Round Robin fashion, if the HOL packet can fit in the frame. Step 2: If the HOL packet of the queue being serviced is too large to fit in the current frame, following Round Robin fashion, find the queue whose HOL packet will fit in the current frame. If no such queue exists, wait for the start of a new frame and schedule the HOL packet in the new frame. Repeat steps 1 and 2. Note: The information (size of the HOL packet in the slave) may be conveyed to the master by slaves using the unused fields in the header of the immediately preceding packet.

Nov Chung-Hsien Hsu16 Look Ahead Round Robin (cont.)  Example 1: Support the number of packets in each queue is n. Queue 1: L 1 = (6, 6, 6, 6, 6, …) Queue 2: L 2 = (8, 8, 8, 8, 8, …) Queue 3: L 3 = (4, 4, 4, 4, 4, …) Queue 4: L 4 = (2, 2, 2, 2, 2, …) Round Robin algorithm : 3n frames. Round Robin with 1-look-ahead algorithm : 2n frames. Performance comparison ratio: 3/2

Nov Chung-Hsien Hsu17 Look Ahead Round Robin (cont.)  Example 2: Support the number of packets in each queue is n. Queue 1: L 1 = (6, 6, 6, 6, 6, …) Queue 2: L 2 = (6, 6, 6, 6, 6, …) …… Queue 2k-1: L 2k-1 = (6, 6, 6, 6, 6, …) Queue 2k: L 2k = (8, 8, 8, 8, 8, …) Queue 2k+1: L 2k+1 = (4, 4, 4, 4, 4, …) Queue 2k+2: L 2k+2 = (4, 4, 4, 4, 4, …) …… Queue 4k-1: L 4k-1 = (4, 4, 4, 4, 4, …) Queue 4k: L 4k = (2, 2, 2, 2, 2, …) Round Robin algorithm : 3kn frames. Round Robin with (2k-1)-look-ahead algorithm : 2kn frames. Performance comparison ratio: 3/2

Nov Chung-Hsien Hsu18 Look Ahead Round Robin (cont.)  Example 3: Support the number of packets in each queue is n. Queue 1: L 1 = (6, 6, 6, 6, 6, …) Queue 2: L 2 = (6, 6, 6, 6, 6, …) …… Queue 2k: L 2k = (6, 6, 6, 6, 6, …) Queue 2k+1: L 2k+1 = (8, 8, 8, 8, 8, …) Queue 2k+2: L 2k+2 = (4, 4, 4, 4, 4, …) Queue 2k+3: L 2k+3 = (4, 4, 4, 4, 4, …) …… Queue 4k+1: L 4k+1 = (4, 4, 4, 4, 4, …) Queue 4k+2: L 4k+2 = (2, 2, 2, 2, 2, …) Round Robin algorithm : (3k+1)n frames. Round Robin with 2k-look-ahead algorithm : (2k+1)n frames. Performance comparison ratio: (3k+1)/(2k+1)

Nov Chung-Hsien Hsu19 Simulation  Special situations –Round Robin with k-look-ahead can reduce the number of frames by 50% over just Round Robin.  General situations: –Turned to simulation.

Nov Chung-Hsien Hsu20 Simulation (cont.)  Environment: –Developed in Java Create pseudo-parallel packet generation and scheduling. Assume packet arrival at the master and the slaves followed a Poisson process. Packet size: 1, 3, 5 Packet be drawn from a uniform distribution.

Nov Chung-Hsien Hsu21 Simulation (cont.)  Variables: –The number of slaves in the piconet. 1 ~ 7 –Packet arrival rate. 0.1 ~ 1.0  The simulation was carried out for 1000 units of times.  Each experiment was repeated 500 times to compute average frame savings.

Nov Chung-Hsien Hsu22 Simulation (cont.) Number of Slaves Average % Savings from RR LARR LA

Nov Chung-Hsien Hsu23 Simulation (cont.) Arrival rates Average % Savings from RR LARR LA

Nov Chung-Hsien Hsu24 Simulation (cont.) Arrival rates Average % Savings from RR LARR LA

Nov Chung-Hsien Hsu25 Simulation (cont.) Arrival rates Average % Savings from RR LARR LA

Nov Chung-Hsien Hsu26 Conclusion  Present two scheduling algorithms for BT. –Look Ahead algorithm –Look Ahead Round Robin algorithm.  LA and LARR perform significantly better than the RR algorithm.

Nov Chung-Hsien Hsu27 Conclusion (cont.)  LARR better than LA –Computational complexity is smaller. –To avoid starvation.  Currently Examining the efficacy of these two algorithms from the power consumption point of view.