Analytical Models for Streaming Media Server Performance Evaluation Qing Wang Minyi Xu May 11, 2001.

Slides:



Advertisements
Similar presentations
Chapter 13 Queueing Models
Advertisements

Waiting Line Management
E&CE 418: Tutorial-4 Instructor: Prof. Xuemin (Sherman) Shen
Module C8 Queuing Economic/Cost Models. ECONOMIC ANALYSES Each problem is different Examples –To determine the minimum number of servers to meet some.
Scalable On-demand Media Streaming Anirban Mahanti Department of Computer Science University of Calgary Canada T2N 1N4.
Prediction-based Prefetching to Support VCR-like Operations in Gossip-based P2P VoD Systems Tianyin Xu, Weiwei Wang, Baoliu Ye Wenzhong Li, Sanglu Lu,
Scalable On-demand Media Streaming with Packet Loss Recovery Anirban Mahanti Department of Computer Science University of Calgary Calgary, AB T2N 1N4 Canada.
INDR 343 Problem Session
Silberschatz, Galvin and Gagne  2002 Modified for CSCI 399, Royden, Operating System Concepts Operating Systems Lecture 19 Scheduling IV.
Lecture 13 – Continuous-Time Markov Chains
Simulation of multiple server queuing systems
Previously Optimization Probability Review Inventory Models Markov Decision Processes.
Simulation A Queuing Simulation. Example The arrival pattern to a bank is not Poisson There are three clerks with different service rates A customer must.
Model Antrian By : Render, ect. Outline  Characteristics of a Waiting-Line System.  Arrival characteristics.  Waiting-Line characteristics.  Service.
Waiting Lines and Queuing Theory Models
Simulation A Queuing Simulation. Example The arrival pattern to a bank is not Poisson There are three clerks with different service rates A customer must.
Analysis of Using Broadcast and Proxy for Streaming Layered Encoded Videos Wilson, Wing-Fai Poon and Kwok-Tung Lo.
Data Broadcast in Asymmetric Wireless Environments Nitin H. Vaidya Sohail Hameed.
© The McGraw-Hill Companies, Inc., Technical Note 6 Waiting Line Management.
Managing Waiting Lines
1 Analysis Of Queues For this session, the learning objectives are:  Learn the fundamental structure of a queueing system.  Learn what needs to be specified.
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.
HHMSM: A Hierarchical Hybrid Multicast Stream Merging Scheme For Large-Scale Video-On-Demand Systems Hai Jin and Dafu Deng Huazhong University of Science.
1 Performance Evaluation of Computer Networks Objectives  Introduction to Queuing Theory  Little’s Theorem  Standard Notation of Queuing Systems  Poisson.
1 Queuing Theory 2 Queuing theory is the study of waiting in lines or queues. Server Pool of potential customers Rear of queue Front of queue Line (or.
Queueing Theory Chapter 17.
Optimal Proxy Cache Allocation for Efficient Streaming Media Distribution Bing Wang, Subhabrata Sen, Micah Adler, and Don Towsley INFOCOM 2002.
Dimensioning the Capacity of True Video-on-Demand Servers Nelson L. S. da Fonseca, Senior Member, IEEE, and Hana Karina S. Rubinsztejn IEEE TRANSACTIONS.
Performance Evaluation of Peer-to-Peer Video Streaming Systems Wilson, W.F. Poon The Chinese University of Hong Kong.
MGTSC 352 Lecture 23: Congestion Management Introduction: Asgard Bank example Simulating a queue Types of congested systems, queueing template Ride’n’Collide.
Little’s Theorem Examples Courtesy of: Dr. Abdul Waheed (previous instructor at COE)
7/3/2015© 2007 Raymond P. Jefferis III1 Queuing Systems.
Queuing Theory. Queuing theory is the study of waiting in lines or queues. Server Pool of potential customers Rear of queue Front of queue Line (or queue)
Model Antrian By : Render, ect. M/M/1 Example 2 Five copy machines break down at UM St. Louis per eight hour day on average. The average service time.
Provisioning Content Distribution Networks for Streaming Media Jussara M. Almeida Derek L. Eager Michael Ferris Mary K. Vernon University of Wisconsin-Madison.
CS Spring 2012 CS 414 – Multimedia Systems Design Lecture 34 – Media Server (Part 3) Klara Nahrstedt Spring 2012.
Lecture 14 – Queuing Systems
ORF Electronic Commerce Spring 2009 April 15, 2009 Week 10 Capacity Analysis Deterministic Model –Assume: each HTTP request takes T r (request time)
Buffer or Suffer Principle
Queuing Networks. Input source Queue Service mechanism arriving customers exiting customers Structure of Single Queuing Systems Note: 1.Customers need.
1 Cache Me If You Can. NUS.SOC.CS5248 OOI WEI TSANG 2 You Are Here Network Encoder Sender Middlebox Receiver Decoder.
Distributing Layered Encoded Video through Caches Authors: Jussi Kangasharju Felix HartantoMartin Reisslein Keith W. Ross Proceedings of IEEE Infocom 2001,
CPSC 441: Multimedia Networking1 Outline r Scalable Streaming Techniques r Content Distribution Networks.
1 1 Slide © 2001 South-Western College Publishing/Thomson Learning Anderson Sweeney Williams Anderson Sweeney Williams Slides Prepared by JOHN LOUCKS QUANTITATIVE.
Segment-Based Proxy Caching of Multimedia Streams Authors: Kun-Lung Wu, Philip S. Yu, and Joel L. Wolf IBM T.J. Watson Research Center Proceedings of The.
4/11: Queuing Models Collect homework, roll call Queuing Theory, Situations Single-Channel Waiting Line System –Distribution of arrivals –Distribution.
Queueing Analysis of Production Systems (Factory Physics)
Analytic Evaluation of Quality of Service for On-Demand Data Delivery Hongfei Guo Haonan Tan
Entities and Objects The major components in a model are entities, entity types are implemented as Java classes The active entities have a life of their.
1 Queueing Theory Frank Y. S. Lin Information Management Dept. National Taiwan University
Queueing Theory What is a queue? Examples of queues: Grocery store checkout Fast food (McDonalds – vs- Wendy’s) Hospital Emergency rooms Machines waiting.
Waiting Lines and Queuing Models. Queuing Theory  The study of the behavior of waiting lines Importance to business There is a tradeoff between faster.
1 Queuing Systems (2). Queueing Models (Henry C. Co)2 Queuing Analysis Cost of service capacity Cost of customers waiting Cost Service capacity Total.
Queueing Theory Dr. Ron Lembke Operations Management.
Chapter 1 Introduction. “Wait-in-line” is a common phenomenon in everywhere. Reason: Demand is more than service. “How long must a customer wait?” or.
Structure of a Waiting Line System Queuing theory is the study of waiting lines Four characteristics of a queuing system: –The manner in which customers.
1 1 Slide Chapter 12 Waiting Line Models n The Structure of a Waiting Line System n Queuing Systems n Queuing System Input Characteristics n Queuing System.
SIMULATION EXAMPLES. Monte-Carlo (Static) Simulation Estimating profit on a sale promotion Estimating profit on a sale promotion Estimating profit on.
© 2015 McGraw-Hill Education. All rights reserved. Chapter 17 Queueing Theory.
Waiting Line Theroy BY, PRAYASH NEUPANE, KARAN CHAND & SANTOSH SHERESTHA.
Random Variables r Random variables define a real valued function over a sample space. r The value of a random variable is determined by the outcome of.
QUEUING THOERY. To describe a queuing system, an input process and an output process must be specified. Examples of input and output processes are: SituationInput.
Simulation of single server queuing systems
March 15, 2001Mark Kalman - ee368c Analysis of Adaptive Media Playout for Stochastic Channel Models Mark Kalman Class Project EE368c.
Chapter 1 Introduction.
Queueing Theory What is a queue? Examples of queues:
Demo on Queuing Concepts
Queuing Systems Don Sutton.
Variability 8/24/04 Paul A. Jensen
Lecture 13 – Queuing Systems
Presentation transcript:

Analytical Models for Streaming Media Server Performance Evaluation Qing Wang Minyi Xu May 11, 2001

Background Streaming media service: Use streams to serve the client requests for large-size, long- duration video-on-demand files Several protocols of such service let client receive data from more than one streams, those streams can then merge to decrease the server bandwidth (stream number). What is merging? Two streams deliver same file to a client, one of them terminates in meantime, the other continues.

Patching and HMSM Patching: client receives data from two streams at the beginning and later two streams merge. The mergee stream continue to finish the full duration. HMSM: client also “listen to” two streams at first, and later two streams also merge, BUT the mergee stream can later merge with other stream (be a merger).

Performance of Patching and HMSM Eager et al: calculate required server bandwidth when patching or HMSM is used. Other system parameters are still needed to analyze performance of patching or HMSM: –waiting time (wait in the queue for service) –balking probability (leave in case of no immediate service) –etc. Measuring the waiting time and balking probability is the goal of this project!

Balking model Simplest Machine Repair Model FCFS center Think node

Balking model Customer in the FCFS of that model: idle streams. S think = average duration of the active streams –get the required stream number using equations from Eager et al. –use Little’s result to get average duration of active streams S FCFS = 1 / (total arrival rate) = inter-arrival time of client requests in stream media service system.

Waiting time Coalescing due to waiting in queue: New-arrived client coalesces with client of the same kind waiting in the queue –coalescing probability AMVA Note: number of one kind of clients waiting in the queue is the same as coalescing probability.

Zipf(  ) function A random variable has a Zipf(  ) distribution if its possibility mass function is: P{X=k} = C / (k 1-  ) The requesting frequency of a certain file is a random variable which suits this distribution.

Experiment – multi-group Evening:total = 125/min (p: probability for choosing this group) Group1: 10 files,  =0.2, T=30min, p=0.4 (CNN News) Group2: 20 files,  =0.3, T=60min, p=0.1 (Badger Herald News) Group3: 25 files,  =0.1, T=100min, p=0.2 (TNT Cable) Group4: 30 files,  =0.25, T=120min, p=0.3 (Starz Cable Movie)

Experiment – multi-group Daytime:total = 125/min (p: probability for choosing this group) Group1: 10 files,  =0.2, T=30min, p=0.6 (CNN News) Group2: 20 files,  =0.3, T=60min, p=0.25 (Badger Herald News) Group3: 25 files,  =0.1, T=100min, p=0.1 (TNT Cable) Group4: 30 files,  =0.25, T=120min, p=0.05 (Startz Cable Movie)

Comparison between evening service pattern and daytime service pattern The long-duration files (movie) are less selected in daytime, short-duration files (news) are more selected in daytime. Pro: longer-duration file less requested; Con: shorter-duration file makes merging less frequently to occur.

Balking Probability vs Server Bandwidth (Optimal patching, multi-group files)

Balking Probability vs Server Bandwidth (HMSM(2,1), multi-group files)

Client Waiting Time vs Server Bandwidth (Optimal patching, multi-group files)

Client Waiting Time vs Server Bandwidth (HMSM(2,1), multi-group files)