June 3, 2015Windows Scheduling Problems for Broadcast System 1 Amotz Bar-Noy, and Richard E. Ladner Presented by Qiaosheng Shi.

Slides:



Advertisements
Similar presentations
Boosting Textual Compression in Optimal Linear Time.
Advertisements

Tight Bounds for Online Class- constrained Packing Hadas Shachnai Bell Labs and The Technion IIT Tami Tamir The Technion IIT.
Routing and Congestion Problems in General Networks Presented by Jun Zou CAS 744.
Shortest Vector In A Lattice is NP-Hard to approximate
Approximation Algorithms Chapter 14: Rounding Applied to Set Cover.
Greedy Algorithms Greed is good. (Some of the time)
Lecture 24 Coping with NPC and Unsolvable problems. When a problem is unsolvable, that's generally very bad news: it means there is no general algorithm.
Approximation, Chance and Networks Lecture Notes BISS 2005, Bertinoro March Alessandro Panconesi University La Sapienza of Rome.
Study Group Randomized Algorithms 21 st June 03. Topics Covered Game Tree Evaluation –its expected run time is better than the worst- case complexity.
A survey of some results on the Firefighter Problem Kah Loon Ng DIMACS Wow! I need reinforcements!
EE 553 Integer Programming
Bounds on Code Length Theorem: Let l ∗ 1, l ∗ 2,..., l ∗ m be optimal codeword lengths for a source distribution p and a D-ary alphabet, and let L ∗ be.
Network Coding in Peer-to-Peer Networks Presented by Chu Chun Ngai
Deterministic Selection and Sorting Prepared by John Reif, Ph.D. Analysis of Algorithms.
A polylogarithmic approximation of the minimum bisection Robert Krauthgamer The Hebrew University Joint work with Uri Feige.
Combinatorial Algorithms
Discrete Structure Li Tak Sing( 李德成 ) Lectures
Optimization of Pearl’s Method of Conditioning and Greedy-Like Approximation Algorithm for the Vertex Feedback Set Problem Authors: Ann Becker and Dan.
Windows Scheduling Problems for Broadcast System 1 Amotz Bar-Noy, and Richard E. Ladner Presented by Qiaosheng Shi.
Approximation Algorithms
Harmonic Broadcasting for Video-on- Demand Service Enhanced Harmonic Data Broadcasting And Receiving Scheme For Popular Video Service Li-Shen Juhn and.
Tirgul 10 Rehearsal about Universal Hashing Solving two problems from theoretical exercises: –T2 q. 1 –T3 q. 2.
Client Buffering Techniques for Scalable Video Broadcasting Over Broadband Networks With Low User Delay S.-H. Gary Chan and S.-H. Ivan Yeung, IEEE Transactions.
1 Adaptive Live Broadcasting for Highly-Demanded Videos Hung-Chang Yang, Hsiang-Fu Yu and Li-Ming Tseng IEEE International Conference on Parallel and Distributed.
3 -1 Chapter 3 The Greedy Method 3 -2 The greedy method Suppose that a problem can be solved by a sequence of decisions. The greedy method has that each.
Data Broadcast in Asymmetric Wireless Environments Nitin H. Vaidya Sohail Hameed.
2002/04/18Chin-Kai Wu, CS, NTHU1 Jitter Control in QoS Network Yishay Mansour and Boaz Patt-Shamir IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 9, NO. 4,
CSE115/ENGR160 Discrete Mathematics 03/03/11 Ming-Hsuan Yang UC Merced 1.
Scalable On-Demand Media Streaming With Packet Loss Recovery Anirban Mahanti, Derek L. Eager, Mary K. Vernon, and David J. Sundaram-Stukel IEEE/ACM Trans.
A general approximation technique for constrained forest problems Michael X. Goemans & David P. Williamson Presented by: Yonatan Elhanani & Yuval Cohen.
Prefix Caching assisted Periodic Broadcast for Streaming Popular Videos Yang Guo, Subhabrata Sen, and Don Towsley.
Optimal Multicast Smoothing of Streaming Video Over the Internet Subhabrata Sen, Don Towsley, Zhi-Li Zhang, and Jayanta K. Dey IEEE J. Selected Areas in.
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley Asynchronous Distributed Algorithm Proof.
Preference Analysis Joachim Giesen and Eva Schuberth May 24, 2006.
Variable-Length Codes: Huffman Codes
Approximation Algorithms Motivation and Definitions TSP Vertex Cover Scheduling.
Admission Control and Dynamic Adaptation for a Proportional-Delay DiffServ-Enabled Web Server Yu Cai.
1 Joint work with Shmuel Safra. 2 Motivation 3 Motivation.
CS Spring 2012 CS 414 – Multimedia Systems Design Lecture 34 – Media Server (Part 3) Klara Nahrstedt Spring 2012.
Max-flow/min-cut theorem Theorem: For each network with one source and one sink, the maximum flow from the source to the destination is equal to the minimal.
Improved results for a memory allocation problem Rob van Stee University of Karlsruhe Germany Leah Epstein University of Haifa Israel WADS 2007 WAOA 2007.
Huffman Codes Message consisting of five characters: a, b, c, d,e
1 Distributed Computing Optical networks: switching cost and traffic grooming Shmuel Zaks ©
Ch. 8 & 9 – Linear Sorting and Order Statistics What do you trade for speed?
© The McGraw-Hill Companies, Inc., Chapter 3 The Greedy Method.
1 Introduction to Approximation Algorithms. 2 NP-completeness Do your best then.
Dynamic Programming.
The Integers. The Division Algorithms A high-school question: Compute 58/17. We can write 58 as 58 = 3 (17) + 7 This forms illustrates the answer: “3.
Maximum Network Lifetime in Wireless Sensor Networks with Adjustable Sensing Ranges Cardei, M.; Jie Wu; Mingming Lu; Pervaiz, M.O.; Wireless And Mobile.
The Selection Problem. 2 Median and Order Statistics In this section, we will study algorithms for finding the i th smallest element in a set of n elements.
NIBEDITA MAULIK GRAND SEMINAR PRESENTATION OCT 21 st 2002.
Restricted Track Assignment with Applications 報告人:林添進.
Princeton University COS 423 Theory of Algorithms Spring 2001 Kevin Wayne Approximation Algorithms These lecture slides are adapted from CLRS.
Competitive Queue Policies for Differentiated Services Seminar in Packet Networks1 Competitive Queue Policies for Differentiated Services William.
Graph Colouring L09: Oct 10. This Lecture Graph coloring is another important problem in graph theory. It also has many applications, including the famous.
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley.
Simulation case studies J.-F. Pâris University of Houston.
The Analysis of Optimal Stream Merging Solutions for Media-on- Demand Amotz Bar-Noy CUNY and Brooklyn College Richard Ladner University of Washington.
1 Windows Scheduling as a Restricted Version of Bin-packing. Amotz Bar-Noy Brooklyn College Richard Ladner Tami Tamir University of Washington.
Scheduling Techniques for Media-on-Demand Amotz Bar-Noy Brooklyn College Richard Ladner Tami Tamir University of Washington.
Rate Distortion Theory. Introduction The description of an arbitrary real number requires an infinite number of bits, so a finite representation of a.
1 Scheduling Techniques for Broadcasting Popular Media. Amotz Bar-Noy Brooklyn College Richard Ladner Tami Tamir University of Washington.
Analysis of Algorithms CS 477/677 Instructor: Monica Nicolescu Lecture 18.
Production systems Scheduling of batch processing.
Data Dissemination and Management (2) Lecture 10
Chapter 5. Optimal Matchings
James B. Orlin Presented by Tal Kaminker
Flow Feasibility Problems
The Selection Problem.
Data Dissemination and Management (2) Lecture 10
Presentation transcript:

June 3, 2015Windows Scheduling Problems for Broadcast System 1 Amotz Bar-Noy, and Richard E. Ladner Presented by Qiaosheng Shi

June 3, 2015Windows Scheduling Problems for Broadcast System 2 Outline Windows scheduling problem  The optimal windows scheduling problem  The optimal harmonic windows scheduling problem Perfect schedule and tree representation Asymptotic bounds The greedy algorithm The combination technique Solutions for small h Open problems & my project plan

June 3, 2015Windows Scheduling Problems for Broadcast System 3 Outline Windows scheduling problem  The optimal windows scheduling problem  The optimal harmonic windows scheduling problem Perfect schedule and tree representation Asymptotic bounds The greedy algorithm The combination technique Solutions for small h Open problems & my project plan

June 3, 2015Windows Scheduling Problems for Broadcast System 4 Windows scheduling problem h slotted channels, and n pages. Each page i has a window size w i. i:1…n  window vector Question: Is there a schedule for the n pages on the h slotted channels  one page each time slot  the gap between two consecutive appearances of page i is no more than w i. The problem is based on the max metric and not the average metric. That is, the next appearance of a page depends only on its previous appearance. But in the average metric, the next appearance of a page depends on all of its previous appearances.

June 3, 2015Windows Scheduling Problems for Broadcast System 5 The push systems application The broadcasting environment consists of:  Clients who wish to access information pages from broadcast channels.  Servers who broadcast the information pages on channels  Providers who supply the information pages Window size of each page (quality of service) is determined by the money providers paid to servers. The server is left with the problem: minimize the number of channels (bandwidth) needed to guarantee the quality of service.  The optimal windows scheduling problem

June 3, 2015Windows Scheduling Problems for Broadcast System 6  Input: A set W={w 1,w 2,…,w n } of requests for broadcasting. A request with window w i needs to be broadcasted at least once in any window of w i time-slots.  Output: A feasible windows scheduling of W.  Goal: minimize number of channels used H(W). Example: Input: W={2,4,5} Output: one channel … There is at least one transmission of in any window of 5 time-slots 5 There is at least one transmission of in any window of 4 time-slots 4 The Optimal Windows scheduling problem H(W)=1

June 3, 2015Windows Scheduling Problems for Broadcast System 7  Medias are broadcast based on customer demand.  A limited number of channels.  The goal: Minimizing clients’ maximal waiting time (delay) with given bandwidth (number of channels).  Assumption  A client that wishes to watch a movie is ‘listening to all the channels’ and is waiting for his movie to start.  Clients have large enough buffer.  Each channel transmits data at the playback rate.  Basic broadcasting schemes  Broadcast popular movies continuously on h channels. The Media-on-Demand application

June 3, 2015Windows Scheduling Problems for Broadcast System 8 Staggered broadcasting [Dan96]: Transmit the movie repeatedly on each of the channels. Guaranteed delay: at most 1/h. The Media-on-Demand application Can we do better? Client’s buffer!

June 3, 2015Windows Scheduling Problems for Broadcast System 9 Partition the movie into segments (or pages). Early segments (or pages) are transmitted more frequently. The Media-on-Demand application The client can start watching the movie without interruptions. Maximal delay: 1/3. arrive watch & buffer 132 (3 pages) Each time-slot has length 1/3. 0 1/3 2/3 1 4/3 5/3 2 C1:C1: C2:C2: … … arrive watch & buffer

June 3, 2015Windows Scheduling Problems for Broadcast System 10 Why does it work? The 1st page is transmitted in any window of one slot. C1:C1: C2:C2: … … The 2nd page is transmitted in any window of two slots. The 3rd page is transmitted at least once in any window of three slots. The Media-on-Demand application

June 3, 2015Windows Scheduling Problems for Broadcast System 11 The movie is partitioned into n pages, 1,..,n. Necessary and sufficient condition: page i is transmitted at least once in any window of i slots (i- window). The client has page i available on time (from his buffer or from the channels). The maximal delay: one slot = 1/n. Therefore, the goal is to maximize n for given h. The Media-on-Demand application  The optimal harmonic windows scheduling problem

June 3, 2015Windows Scheduling Problems for Broadcast System 12  Given h, maximize n such that each i in 1,..,n is scheduled at least once in i time slots. The maximum n is denoted by N(h). The optimal harmonic windows scheduling problem … … … C1C1 C2C2 C3C3 111 … C1C1 Examples: h=1, n=1, N(1)=1 C1C C2C2 … … h=2, n=3. N(2)=3 h=3, n=9. N(3)=9?

June 3, 2015Windows Scheduling Problems for Broadcast System 13 Perfect channel schedule Channel schedule: each page is scheduled on a single channel. A schedule S is called cyclic if it is an infinite concatenation of a finite sequence. Another definition: Matrix schedule

June 3, 2015Windows Scheduling Problems for Broadcast System 14 Perfect channel schedule Perfect channel schedule: For page i, there exists a, page i gets one time slot exactly every w i ’ time slots.  the window size of page i in the perfect channel schedule.  Perfect channel schedule is cyclic (least common multiple). Several points:  Avoid busy-waiting: the client actively listen until its movie arrives.  Not optimal for windows scheduling problem  Finding an optimal perfect channel schedule is NP-hard in general.  Only need to record three numbers for one page: channel number, period length and offset.

June 3, 2015Windows Scheduling Problems for Broadcast System 15 Tree representation … … … C1C1 C2C2 C3C3 Page Channel Period Offset Tree is simple

June 3, 2015Windows Scheduling Problems for Broadcast System 16 Tree representation One ordered tree per channel Leaves represent the pages Offset The period of each page is the product of the degrees of the nodes on the path from the root to its corresponding leaf.

June 3, 2015Windows Scheduling Problems for Broadcast System 17 Tree representation PageABCD Period2666 Offset

June 3, 2015Windows Scheduling Problems for Broadcast System 18 Tree representation PageABCDEFGH Period Offset

June 3, 2015Windows Scheduling Problems for Broadcast System 19 If all leaves are distinct in forest, the corresponding schedule is perfect channel schedule. Tree representation However, there exist perfect channel schedule that cannot be embedded in a tree. ABCD1D2D3AD4D5D6D7BAD8D9 D10D11CAD12D13BD14D15AD16D17D18D19D20 … Can we always construct an ordered tree for a perfect channel schedule? Can we always get the perfect channel schedule from an ordered tree? AAA AA BB BC C D1D2D3D4D5D6D7D8D9 D10D11D12D13D14D15D16D17D18D19D20 Degree of root must divide the periods 6, 10, 15

June 3, 2015Windows Scheduling Problems for Broadcast System 20 Asymptotic bounds for H(W) Page i requires at least a fraction of a channel Upper bound  It is achieved by constructing a perfect channel schedule. Lower bound  For any window vector Minimum number of channels needed to schedule window vector W, N(h)

June 3, 2015Windows Scheduling Problems for Broadcast System 21 Upper bound for H(W) --- simple case Window sizes are all powers of 2. Lemma: there exists a perfect schedule that uses exactly channels. (the first lemma) First upper bound (round the window sizes down to the nearest power of 2): For any window vector W, there exists a perfect schedule that uses no more than channels.

June 3, 2015Windows Scheduling Problems for Broadcast System 22 Upper bound for H(W) --- simple case All the window sizes are powers of 2 multiplied by some number u. Lemma: If all the are of the form for some and, then there exists a perfect schedule that uses exactly channels. (the 2nd lemma) Construct an algorithm that for given window vector W creates perfect schedules with about channels.

June 3, 2015Windows Scheduling Problems for Broadcast System 23 Upper bound for H(W) --- the algorithm The algorithm use two parameters k and x that are optimized to obtain the best bound. k: the depth of the recursion x: is optimized for each value of k. If k=1, round window size down to closest to get a schedule with at most channels. If k>1, partition the window vector W into x vectors denoted by.  is rounded down to maximal such that, is an odd number and for some u. Then  The set of such that is denoted by

June 3, 2015Windows Scheduling Problems for Broadcast System 24 Upper bound for H(W) --- the algorithm channels needed to schedule all windows in Some windows scheduled into non-fully used channels. The set of all these windows is denoted by If x is larger, then is closer to. However, is too big. If x is smaller, then is small. But is too small compared to For each k, find the best value for x.

June 3, 2015Windows Scheduling Problems for Broadcast System 25 Upper bound for H(W) --- example Let W= At least 3 channels to schedule the windows in W (2<h(W)<3) k=1: W ’ = k=2, x=3: We get the following 3 vectors 3 4 K=1

June 3, 2015Windows Scheduling Problems for Broadcast System 26 Upper bound for H(W) --- major lemma Define r as mapping from the positive integer to the reals by: for k=1 for k>1 The function r is monotonic increasing function whose exist and is approximately For window vector W and positive integers k if then there exists a perfect schedule with number of channels bounded above by

June 3, 2015Windows Scheduling Problems for Broadcast System 27 Upper bound for H(W) --- major lemma Theorem: Every window vector W, with h(W)>1, has perfect schedule using number of channels bounded above by, where Theorem: For any window vector W, there exists an algorithm for the optimal windows scheduling problem yielding a solution that is within a factor of of the optimal solution.

June 3, 2015Windows Scheduling Problems for Broadcast System 28 Bounds on N(h) Lower bound of H(W):  Upper bound of H(W):  Given h channels, maximize n such that each page i is scheduled at least once in any consecutive i slots

June 3, 2015Windows Scheduling Problems for Broadcast System 29 Outline Windows scheduling problem  The optimal windows scheduling problem  The optimal harmonic windows scheduling problem Perfect schedule and tree representation Asymptotic bounds The greedy algorithm The combination technique Solutions for small h Open problems & my project plan