Measurement and Estimation of Network QoS among Peer Xbox Game Players Youngki Lee, KAIST Sharad Agarwal, Microsoft Research Chris Butcher, Bungie Studio.

Slides:



Advertisements
Similar presentations
Colyseus: A Distributed Architecture for Online Multiplayer Games
Advertisements

Ch. 12 Routing in Switched Networks Routing in Packet Switched Networks Routing Algorithm Requirements –Correctness –Simplicity –Robustness--the.
Ningning HuCarnegie Mellon University1 Optimizing Network Performance In Replicated Hosting Peter Steenkiste (CMU) with Ningning Hu (CMU), Oliver Spatscheck.
Characterizing Residential Broadband Networks Marcel Dischinger †, Andreas Haeberlen †‡, Krishna P. Gummadi †, Stefan Saroiu* † MPI-SWS, ‡ Rice University,
Switchboard: A Matchmaking System for Multiplayer Mobile Games Justin Manweiler, Sharad Agarwal, Ming Zhang, Romit Roy Choudhury, Paramvir Bahl ACM MobiSys.
W. Feng, “A Long-term Study of a Popular MMORPG", NetGames 2007, Sept , A Long-term Study of a Popular MMORPG Wu-chang Feng Debanjan Saha David.
On the Geographic Distribution of On- line Game Servers and Players Wu-chang FengWu-chi Feng Presented By: Abhishek Gupta.
Improving Online Gaming Quality using Detour Paths Cong Ly, Cheng-Hsin Hsu, and Mohamed Hefeeda Simon Fraser University, Canada Deutsche Telekom Labs,
Wide-scale Botnet Detection and Characterization Anestis Karasaridis, Brian Rexroad, David Hoeflin.
Multiple constraints QoS Routing Given: - a (real time) connection request with specified QoS requirements (e.g., Bdw, Delay, Jitter, packet loss, path.
On the Geographic Distribution of On- line Game Servers and Players Wu-chang FengWu-chi Feng Discussion moderated By: John Carter.
Network Characteristics for Server Selection in Online Games Mark Claypool Computer Science Department Worcester Polytechnic Institute Worcester, Massachusetts,
1 Network Tomography Venkat Padmanabhan Lili Qiu MSR Tab Meeting 22 Oct 2001.
Characterization by Measurement of a CDMA 1x EVDO Network Presenter: Mingzhe Li Wireless Internet Conference (WICON’06) Boston, Massachusetts,
A Hierarchical Characterization of a Live Streaming Media Workload E. Veloso, V. Almeida W. Meira, A. Bestavros, S. Jin Proceedings of Internet Measurement.
Who Talks to Whom: Using BGP Data for Scaling Interdomain Resource Reservation Ping Pan and Henning Schulzrinne Columbia University ISMA Workshop – Leiden,
Delayed Internet Routing Convergence Craig Labovitz, Abha Ahuja, Abhijit Bose, Farham Jahanian Presented By Harpal Singh Bassali.
Available bandwidth measurement as simple as running wget D. Antoniades, M. Athanatos, A. Papadogiannakis, P. Markatos Institute of Computer Science (ICS),
T. S. Eugene Ng Mellon University1 Global Network Positioning: A New Approach to Network Distance Prediction Tze Sing Eugene.
Impact of BGP Dynamics on Intra-Domain Traffic Patterns in the Sprint IP Backbone Sharad Agarwal, Chen-Nee Chuah, Supratik Bhattacharyya, Christophe Diot.
1 Measurement-based Characterization of a Collection of On-line Games Chris Chambers Wu-chang Feng Portland State University Sambit Sahu Debanjan Saha.
Ningning HuCarnegie Mellon University1 A Measurement Study of Internet Bottlenecks Ningning Hu (CMU) Joint work with Li Erran Li (Bell Lab) Zhuoqing Morley.
End-to-End Issues. Route Diversity  Load balancing o Per packet splitting o Per flow splitting  Spill over  Route change o Failure o policy  Route.
Amir Rasti Daniel Stutzbach Reza Rejaie The ION P2P Project University of Oregon On the Long-term Evolution of the Two-Tier.
A Traffic Characterization of Popular On-Line Games Wu-Chang Feng, Francis Chang, Wu- Chi Feng, and Jonathan Walpole IEEE/ACM Trans. Networking, Jun
Measuring ISP topologies with Rocketfuel Ratul Mahajan Neil Spring David Wetherall University of Washington ACM SIGCOMM 2002.
The Effects of Latency on User Performance in Warcraft III Nathan Sheldon, Eric Gerard, Seth Borg, Mark Claypool, Emmanuel Agu Computer Science Department.
Network Analysis of Counter-strike and Starcraft Mark Claypool, David LaPoint, Josh Winslow Worcester Polytechnic Institute Worcester, MA, USA
Submission doc.: IEEE /0061r6 January 2015 Allan Jones, ActivisionSlide 1 FPS Network Traffic Model Date: Authors:
Characterizing Residential Broadband Networks Marcel Dischinger †, Andreas Haeberlen †‡, Krishna P. Gummadi †, Stefan Saroiu* † MPI-SWS, ‡ Rice University,
1 Efficient Management of Data Center Resources for Massively Multiplayer Online Games V. Nae, A. Iosup, S. Podlipnig, R. Prodan, D. Epema, T. Fahringer,
KaZaA: Behind the Scenes Shreeram Sahasrabudhe Lehigh University
Presentation by Manasee Conjeepuram Krishnamoorthy.
1 Measurement-based Characterization of a Collection of On-line Games Chris Chambers Wu-chang Feng Portland State University Sambit Sahu Debanjan Saha.
Ao-Jan Su, David R. Choffnes, Fabián E. Bustamante and Aleksandar Kuzmanovic Department of EECS Northwestern University Relative Network Positioning via.
PIC: Practical Internet Coordinates for Distance Estimation Manuel Costa joint work with Miguel Castro, Ant Rowstron, Peter Key Microsoft Research Cambridge.
End-to-end QoE Optimization Through Overlay Network Deployment Bart De Vleeschauwer, Filip De Turck, Bart Dhoedt and Piet Demeester Ghent University -
Skype P2P Kedar Kulkarni 04/02/09.
Advanced Networking Lab. Given two IP addresses, the estimation algorithm for the path and latency between them is as follows: Step 1: Map IP addresses.
11 Experimental and Analytical Evaluation of Available Bandwidth Estimation Tools Cesar D. Guerrero and Miguel A. Labrador Department of Computer Science.
Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 1 Efficient Path Aggregation and Error Control for Video Streaming OMESH TICKOO, Shiv Kalyanaraman,
1 A Framework for Measuring and Predicting the Impact of Routing Changes Ying Zhang Z. Morley Mao Jia Wang.
Submission doc.: IEEE /0061r5 January 2015 Allan Jones, ActivisionSlide 1 FPS Network Traffic Model Date: Authors:
Determining the Geographic Location of Internet Hosts Venkata N. Padmanabhan Microsoft Research Lakshminarayanan Subramanian University of California at.
N. Hu (CMU)L. Li (Bell labs) Z. M. Mao. (U. Michigan) P. Steenkiste (CMU) J. Wang (AT&T) Infocom 2005 Presented By Mohammad Malli PhD student seminar Planete.
Networked Games Objectives – –Understand the types of human interaction that a network game may provide and how this influences game play. –Understand.
A Bandwidth Estimation Method for IP Version 6 Networks Marshall Crocker Department of Electrical and Computer Engineering Mississippi State University.
Requirement Analysis Guidelines. 2 Process Model for Requirement Analysis Gather Requirements Develop Service Metrics To measure performance Characterizing.
KAIS T On the problem of placing Mobility Anchor Points in Wireless Mesh Networks Lei Wu & Bjorn Lanfeldt, Wireless Mesh Community Networks Workshop, 2006.
JEHN-RUEY JIANG, GUAN-YI SUNG, JIH-WEI WU NATIONAL CENTRAL UNIVERSITY, TAIWAN PRESENTED BY PROF. JEHN-RUEY JIANG LOM: A LEADER ORIENTED MATCHMAKING ALGORITHM.
1 Network Tomography Using Passive End-to-End Measurements Venkata N. Padmanabhan Lili Qiu Helen J. Wang Microsoft Research DIMACS’2002.
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.
1 NetProfiler: Profiling Networks From the Edge Venkat Padmanabhan Microsoft Research June 2005 With Sharad Agarwal (MSR), Jitu Padhye (MSR), Dilip Joseph.
Patch Scheduling for On-line Games Chris Chambers Wu-chang Feng Portland State University.
Predicting the Perceived Quality of a First Person Shooter Game The Team Fortress 2 T-Model David Dwyer Eric Finn Advisor: Mark Claypool 1.
10-Year History of Internet Delay 1 April 24, 2010, DK Lee, Kenjiro Cho*, Gianluca Iannaccone**, Sue Moon CAIDA-WIDE-CASFI Joint Workshop.
Proposal Presentation Inferring Geographic proximity of the Internet Node using a Compound Metric M. Kamran Nishat.
1 Flow-Aware Networking Introduction Concepts, graphics, etc. from Guide to Flow-Aware Networking: Quality-of-Service Architectures and Techniques for.
Drafting Behind Akamai (Travelocity-Based Detouring) Ao-Jan Su, David R. Choffnes, Aleksandar Kuzmanovic and Fabián E. Bustamante Department of EECS Northwestern.
PlanetSeer: Internet Path Failure Monitoring and Characterization in Wide-Area Services Ming Zhang, Chi Zhang Vivek Pai, Larry Peterson, Randy Wang Princeton.
Vivaldi: A Decentralized Network Coordinate System
FPS Network Traffic Model
Early Measurements of a Cluster-based Architecture for P2P Systems
Phillipa Gill University of Toronto
FPS Network Traffic Model
Measurement-based Characterization of a Collection of On-line Games
FPS Network Traffic Model
Transport Layer Identification of P2P Traffic
End-to-End Internet Delay Behavior
FPS Network Traffic Model
Presentation transcript:

Measurement and Estimation of Network QoS among Peer Xbox Game Players Youngki Lee, KAIST Sharad Agarwal, Microsoft Research Chris Butcher, Bungie Studio Jitu Padhye, Microsoft Research

A series of online multiplayer game via Xbox Live ▫ First Person Shooter (FPS) game ▫ 15 million copies sold worldwide We focus on Halo 3 for data collection and analysis. ▫ Halo 3 has a large set of widely distributed player population. ▫ released on September 25,

P2P architecture of Halo 3  Xbox console

P2P architecture of Halo P2P, a peer as a server  Xbox console Xbox Live matchmaking service Network QoS between the server peer and other client peers is important for game quality. ▫ excellent experience: latency (< 50ms), BW (50~70Kbps). ▫ minimum requirement: latency ( 30Kbps).

QoS probing among peers Xbox Live matchmaking service Query: Give me a list of hosts that satisfy my criteria Probing using the packet-pair technique 5 Candidate hosts

Motivation Understand network path quality (NPQ) among peer game players and characteristics of the players ▫ NPQ in terms of network delay and capacity Address the problem of NPQ measurement overhead ▫ improve user pre-game experience  probe fewer, better candidate hosts Limited publications on large-scale E2E network characterization ▫ Planetlab-based end-to-end NPQ studies: O(100) nodes ▫ king-based end-to-end NPQ studies O(1000) nodes ▫ several studies of provisioned server based games 6

Methodology 1.Collect probe data among peer game players a)consoles report the probe results back to Xbox live service. 2.Understand characteristics of peer game playing 3.Understand NPQ between peer game players 4.Examine stability and predictability of NPQ a)propose three simple predictors  IP history, prefix history, geography b)examine robustness of the predictors 7

Outline Background Motivation Analysis on probe data ▫ general characteristics ▫ NPQ results NPQ prediction ▫ IP history predictor ▫ prefix history predictor ▫ geography predictor Conclusion 8

Data Session data (per game attempted) ▫ time, session-id, src IP NPQ measurement data (per probing to a host) ▫ session-id, dest IP ▫ # of packet-pairs sent, # of packet-pairs rcvd ▫ minimum and median latency ▫ average downstream and upstream capacity Player locations calculated from their IP addresses ▫ MaxMind database provides mapping between locations and IP addresses 9

Basic statistics million probes among 5.6 million IP addresses !!!  (50 days) sessions distinct IPs total probes 39,803,350 5,658, ,085,887

Geographic distribution 11 85% in USA 13% in Europe 2% in Asia, Australia

Player characterization Strong diurnal pattern (peaks between 2 ~ 8PM, UTC time) Most players played a few games, only some a lot Probe distribution per game trial (session) ▫ 90% of sessions probed fewer than 10 hosts, but some a lot

Delay distribution 25% of the delay measurement are above 150ms. ▫ 150 ms: upper bound for responsive experience in FPS games %

Capacity distribution Peaks around typical broadband capacities in USA. ▫ marginal error due to the packet pair technique. 14

Outline Background Motivation Analysis on probe data ▫ general characteristics ▫ NPQ results NPQ prediction ▫ IP history predictor ▫ prefix history predictor ▫ geography predictor Conclusion 15

Predictors Predict NPQ without probing ▫ to disqualify a host, select a host, do quick re-probe ▫ potentially reduce the user-wait time and probe traffic IP/Prefix history predictor ▫ reuse the previous probe results between the same IP pair ▫ reuse results between two peers within the same prefix pair  determine prefixes by BGP table (12/27/2007 RouteViews) Geography predictor ▫ predict delay or capacity based on the geographic distance 16

IP history predictor (delay) Delays are very consistent over time, even for 50 days ▫ excellent predictor for delay 17 CV= Stdev/Mean, small CV = small variation (50 days)

IP history predictor (capacity) Capacities are also quite consistent over time. ▫ decent predictor for downstream capacity 18 (50 days)

Prefix history predictor Quite consistent, but more variation compared to IP pairs ▫ outliers mostly caused the variation. ▫ good predictor for delay after removing outliers. 19 (50 days)

Geography predictor Distance has strong correlation with minimum delay ▫ good predictor for removing hosts with high latency 20 Distance (miles) Delay (ms)

Conclusions Large-scale end host latency and capacity characterization Large-scale P2P game network characterization ▫ 126 million probes among 5.6 million unique IPs NPQ prediction for delay ▫ IP history : great ! ▫ prefix history: good after removing outliers ▫ geography : great for removing distant hosts NPQ prediction for capacity ▫ IP history: decent! ▫ prefix history: not feasible ▫ geography: not feasible 21