A Hierarchical Characterization of a Live Streaming Media Workload IEEE/ACM Trans. Networking, Feb. 2006 Eveline Veloso, Virg í lio Almeida, Wagner Meira,

Slides:



Advertisements
Similar presentations
Watching Television Over an IP Network & TV-Watching Behavior Research presented by Weiping He.
Advertisements

Waiting Line Management
Modeling User Activities in a Large IPTV System Tongqing Qiu, Jun (Jim) Xu (Georgia Tech) Zihui Ge, Seungjoon Lee, Jia Wang, Qi Zhao (AT&T Lab – Research)
Scalable On-demand Media Streaming Anirban Mahanti Department of Computer Science University of Calgary Canada T2N 1N4.
Scalable On-demand Media Streaming with Packet Loss Recovery Anirban Mahanti Department of Computer Science University of Calgary Calgary, AB T2N 1N4 Canada.
L-21 Multicast. L -15; © Srinivasan Seshan, Overview What/Why Multicast IP Multicast Service Basics Multicast Routing Basics DVMRP Overlay.
Computer Science Generating Streaming Access Workload for Performance Evaluation Shudong Jin 3nd Year Ph.D. Student (Advisor: Azer Bestavros)
Adapted from Menascé & Almeida.1 Workload Characterization for the Web.
Measurement, Modeling, and Analysis of a Peer-2-Peer File-Sharing Workload Presented For Cs294-4 Fall 2003 By Jon Hess.
2005/2/23 HUT T Characterizing Web Workload of Mobile Clients Chuang Yu Juha Raitio.
1 10 Web Workload Characterization Web Protocols and Practice.
GlobeTraff A traffic workload generator for the performance evaluation of ICN architectures K.V. Katsaros, G. Xylomenos, G.C. Polyzos A.U.E.B. (presented.
1 Network Traffic Measurement and Modeling Carey Williamson Department of Computer Science University of Calgary.
An Empirical Study of Real Audio Traffic A. Mena and J. Heidemann USC/Information Sciences Institute In Proceedings of IEEE Infocom Tel-Aviv, Israel March.
On the Geographic Location of Internet Resources CSCI 780, Fall 2005.
Multi-level Application-based Traffic Characterization in a Large-scale Wireless Network Maria Papadopouli 1,2 Joint Research with Thomas Karagianis 3.
Analysis of Using Broadcast and Proxy for Streaming Layered Encoded Videos Wilson, Wing-Fai Poon and Kwok-Tung Lo.
Distributed Servers Architecture for Networked Video Services S.-H. Gary Chan and Fouad Tobagi Presented by Todd Flanagan.
On the Geographic Distribution of On- line Game Servers and Players Wu-chang FengWu-chi Feng Discussion moderated By: John Carter.
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 Hierarchical Characterization of a Live Streaming Media Workload E. Veloso, V. Almeida W. Meira, A. Bestavros, S. Jin Proceedings of Internet Measurement.
OS Fall ’ 02 Performance Evaluation Operating Systems Fall 2002.
Network Traffic Measurement and Modeling CSCI 780, Fall 2005.
Distributed Servers Architecture for Networked Video Services S. H. Gary Chan, Member IEEE, and Fouad Tobagi, Fellow IEEE.
Data Communication and Networks Lecture 13 Performance December 9, 2004 Joseph Conron Computer Science Department New York University
1 Can Internet Video-on-Demand be Profitable? Cheng Huang, Jin Li (Microsoft Research Redmond), Keith W. Ross (Polytechnic University) ACM SIGCOMM 2007.
Copyright © 2005 Department of Computer Science CPSC 641 Winter Network Traffic Measurement A focus of networking research for 20+ years Collect.
1 Lecture 24: Interconnection Networks Topics: communication latency, centralized and decentralized switches (Sections 8.1 – 8.5)
Web Caching Robert Grimm New York University. Before We Get Started  Illustrating Results  Type Theory 101.
Graphs and Topology Yao Zhao. Background of Graph A graph is a pair G =(V,E) –Undirected graph and directed graph –Weighted graph and unweighted graph.
A Hybrid Caching Strategy for Streaming Media Files Jussara M. Almeida Derek L. Eager Mary K. Vernon University of Wisconsin-Madison University of Saskatchewan.
7/3/2015© 2007 Raymond P. Jefferis III1 Queuing Systems.
OS Fall ’ 02 Performance Evaluation Operating Systems Fall 2002.
A Traffic Characterization of Popular On-Line Games Wu-Chang Feng, Francis Chang, Wu- Chi Feng, and Jonathan Walpole IEEE/ACM Trans. Networking, Jun
Internet Queuing Delay Introduction How many packets in the queue? How long a packet takes to go through?
CS 6401 Network Traffic Characteristics Outline Motivation Self-similarity Ethernet traffic WAN traffic Web traffic.
Can Internet Video-on-Demand Be Profitable? SIGCOMM 2007 Cheng Huang (Microsoft Research), Jin Li (Microsoft Research), Keith W. Ross (Polytechnic University)
Add image. 3 “ Content is NOT king ” today 3 40 analog cable digital cable Internet 100 infinite broadcast Time Number of TV channels.
Self-Similarity of Network Traffic Presented by Wei Lu Supervised by Niclas Meier 05/
Traffic Modeling.
Can Internet VoD be Profitable? Cheng Huang (MSR), Jin Li (MSR), Keith W. Ross (NY Polytechnique)
1 An SLA-Oriented Capacity Planning Tool for Streaming Media Services Lucy Cherkasova, Wenting Tang, and Sharad Singhal HPLabs,USA.
An Analysis of WoW Players’ Game Hours Matt Ross, Christian Ebinger, Anthony Morgan.
A Geographical Characterization of YouTube: a Latin American View Fernando Duarte, Fabrício Benevenuto, Virgílio Almeida, Jussara Almeida Federal University.
Web Cache Replacement Policies: Properties, Limitations and Implications Fabrício Benevenuto, Fernando Duarte, Virgílio Almeida, Jussara Almeida Computer.
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.
Spatio-Temporal Modeling of Traffic Workload in a Campus WLAN Felix Hernandez-Campos 3 Merkouris Karaliopoulos 2 Maria Papadopouli 1,2,3 Haipeng Shen 2.
Improving MBMS Security in 3G Wenyuan Xu Rutgers University.
Understanding the Performance of Web Caching System with an Analysis Model and Simulation Xiaosong Hu Nur Zincir-Heywood Sep
Measuring the Congestion Responsiveness of Internet Traffic Ravi Prasad & Constantine Dovrolis Networking and Telecommunications Group College of Computing.
Microsoft Research1 Characterizing Alert and Browse Services for Mobile Clients Atul Adya, Victor Bahl, Lili Qiu Microsoft Research USENIX Annual Technical.
TEL500-Voice Communications SIP-based VoIP Traffic Behavior Profiling and Its Application Devesh Mendiratta & Sameer Deshmukh MS-Telecommunication State.
Internet Measurment Multimedia 1. Properties Challenges Tools State of the Art 2.
Pin-Yun Tarng / An Analysis of WoW Players’ Game Hours Network and Systems Laboratory nslab.ee.ntu.edu.tw IEEE/IFIP DSN 2008 Network and Systems Laboratory.
Ch 6. Multimedia Networking Myungchul Kim
NTMS 2012 GlobeTraff: a traffic workload generator for the performance evaluation of future Internet architectures K.V. Katsaros, G. Xylomenos, G.C. Polyzos.
Patch Scheduling for On-line Games Chris Chambers Wu-chang Feng Portland State University.
Doc.: IEEE /1263r2 Submission Dec 2009 Z. Chen, C. Zhu et al [Preliminary Simulation Results on Power Saving] Date: Authors: Slide.
Video Caching in Radio Access network: Impact on Delay and Capacity
#16 Application Measurement Presentation by Bobin John.
1 Internet Traffic Measurement and Modeling Carey Williamson Department of Computer Science University of Calgary.
Data Network Design and Evaluation Dr Usman Saeed Assistant Professor Faculty of Computing and Information Technology North Jeddah Branch King Abdulaziz.
Ó 1998 Menascé & Almeida. All Rights Reserved.1 Part VIII Web Performance Modeling (Book, Chapter 10)
Does Internet media traffic really follow the Zipf-like distribution? Lei Guo 1, Enhua Tan 1, Songqing Chen 2, Zhen Xiao 3, and Xiaodong Zhang 1 1 Ohio.
March 15, 2001Mark Kalman - ee368c Analysis of Adaptive Media Playout for Stochastic Channel Models Mark Kalman Class Project EE368c.
Implementing a Load-balancing Web Server Using Red Hat Cluster Suite
Web switch support for differentiated services
Network Traffic Modeling
Prestented by Zhi-Sheng, Lin
Waiting Line Models Waiting takes place in virtually every productive process or service. Since the time spent by people and things waiting in line is.
Presentation transcript:

A Hierarchical Characterization of a Live Streaming Media Workload IEEE/ACM Trans. Networking, Feb Eveline Veloso, Virg í lio Almeida, Wagner Meira, Jr., Azer Bestavros, and Shudong Jin

Motivation  The characteristics of live media and stored media are different. Stored media object: user driven  Be directly influenced by user preferences Live media object: content driven  Be directly influenced by aspects related to the nature of the object A Traffic Characterization of Popular On-Line Games: 203/A_Traffic_Characterization_of_Popular_On-Line_Games.ppt

Basic statistics of the trace used in this paper Microsoft Media Server … stream 1 stream 2 48 different cameras 7 Kbps 18 Kbps 32 Kbps 57 Kbps

Characterization hierarchy  Client layer  Session layer The interval of time during which the client is actively engaged in requesting live streams that are part of the same service such that the duration of any period of no transfers between the server and the client does not exceed a preset threshold T off.  Transfer layer In session ON time During transfer ON time, a client is served one or more live streams. Transfer OFF times correspond loosely to “ think ” times.

Relationship between client activities and ON/OFF times

Client layer characteristics  Topological and geographical distribution of client population Zipf-like distribution  Most requests are issued from a few regions  Client concurrency profile  Client interarrival times  Client interest profile

Client diversity: IP addresses over ASs Autonomous System (AS): the unit of router policy, either a single network or a group of networks that is controlled by a common network administrator

Client diversity: transfers over ASs

Client diversity: transfers over countries

Client layer characteristics  Topological and geographical distribution of client population  Client concurrency profile Periodic behavior  Client interarrival times  Client interest profile

Cumulative distribution of number of active clients (cumulative)

Temporal behavior of number of active clients: over entire trace

Temporal behavior of number of active clients: daily Weekend have slightly higher clients than weekdays

Temporal behavior of number of active clients: hourly

Client layer characteristics  Topological and geographical distribution of client population  Client concurrency profile  Client interarrival times Pareto distribution Piece-wise-stationary Poisson process  Client interest profile

Client interarrival times: frequency What is the unit of frequency? It might be 1.instance/second (x) 2.instance/request (?) 3.percentage (?)

Client interarrival times: CCDF CCDF: Complementary Cumulative Distribution Function

Discuss  The client arrival process is not stationary in that it is highly dependent on time.  It is natural to assume that over a very short time interval, such a process would be stationary, and may indeed be Poisson. Piece-wise-stationary Poisson arrival  15 min.

Client interarrival times: piece-wise- stationary Poisson process

Client layer characteristics  Topological and geographical distribution of client population  Client concurrency profile  Client interarrival times  Client interest profile Characterizing live content popularity is not meaningful  characterizing the “ interest ” of a client in the live content is more appropriate Zipf-like distribution  Most requests are issued from a few clients

Client interest profile: client rank v.s. transfer frequency Rank: number of transfers for that client

Client interest profile: client rank v.s. session frequency Rank: number of sessions for that client

Session layer characteristics  Number of sessions Threshold T off  Session ON time  Session OFF time  Transfers per session  Interarrivals of session transfers

Relationship between number of sessions and T off 3600

Session layer characteristics  Number of sessions  Session ON time Lognormal distribution  Session OFF time  Transfers per session  Interarrivals of session transfers

Distribution of session ON times

Session layer characteristics  Number of sessions  Session ON time  Session OFF time Exponential distribution  Transfers per session  Interarrivals of session transfers

Distribution of session OFF times

Session layer characteristics  Number of sessions  Session ON time  Session OFF time  Transfers per session Pareto distribution  Interarrivals of session transfers

Number of transfers per session: frequency

Number of transfers per session: CCDF

Session layer characteristics  Number of sessions  Session ON time  Session OFF time  Transfers per session  Interarrivals of session transfers Lognormal distribution

Session transfer interarrivals: frequency

Transfer layer characteristics  Number of concurrent transfers Exponential distribution  Transfer length and client stickiness  Transfer interarrivals  Transfer bandwidth

Concurrent transfers over all sessions (cumulative)

Transfer layer characteristics  Number of concurrent transfers  Transfer length and client stickiness Lognormal distribution  The long tail of the transfer length distribution is due to the client ’ s willingness to “ stick ” to the live stream.  Transfer interarrivals  Transfer bandwidth

Transfer lengths

Transfer layer characteristics  Number of concurrent transfers  Transfer length and client stickiness  Transfer interarrivals Like client arrivals Pareto distribution  Transfer bandwidth

Transfer interarrival times

Temporal behavior of transfer interarrival times: over entire trace

Temporal behavior of transfer interarrival times: daily Weekends have lower average interarrivals than weekdays (but more clients)  Due to channel browsing?

Temporal behavior of transfer interarrival times: hourly

Transfer layer characteristics  Number of concurrent transfers  Transfer length and client stickiness  Transfer interarrivals  Transfer bandwidth Client-bound bandwidth Congestion-bound bandwidth

Aggregate bandwidth

Frequency distributions of transfer bandwidth client: 58.6 Kbps 32.5 Kbps 17.6 Kbps 6.87 Kbps congestion

Across multiple live media workloads  Another live streaming server for a “ news and sports ” radio station  The differences of two live streaming services Client interarrival times Session transfer interarrival times Transfer interarrival times  These differences are due to the different interactions between clients and live streams in the workloads.

Summary of the characteristics of the “Reality Show” and “News and Sports”