1. Internet Measurement Online Games 1

2. Why Online Games? One of the fastest growing areas of the Internet More recently non-sequential and interactive gaming has become popular On-line games are big business  60% of all Americans play video games Hosting games very costly (30% of revenue) 2

3. Online Games Games and Measurement Properties Networked Games Measurement Challenges State of the Art 3

4. Online Games Networked Games Measurement Challenges State of the Art 4

5. Game Measurement Properties Different genres:  First Person Shooters (FPS) Most popular type of online gaming High real time requirements  Real Time Strategy (RTS)  Massive Multiplayer Online Role Playing Games (MMORPGs) Different Platforms: XBOX, PS, Nintendo 5

6. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 6

7. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 7

8. Traffic Characterization First property: Fraction of Internet traffic and individual popularity of games Sample traffic flowing to and from port numbers of games (like P2P) Individual game characterization  Traditional metrics like size, inter-arrival time of packets  Behavioural differences between clients and server  Large amount of games trafiic take place over proprietary networks – surveys used in these cases. 8

9. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 9

10. Online Games Games and Measurement Properties Game System Architecture CentralizedDecentralizedHybrid Networked Games Measurement Challenges State of the Art 10

11. Online Games Games and Measurement Properties Game System Architecture DecentralizedHybrid Networked Games Measurement Challenges State of the Art 11

12. Centralized Architecture 1 All interaction requests sent through a central server All clients not required to know movements of all other clients at any given instant Server decides what each client needs to know 12

13. Centralized Architecture 2 Server requirements: High processing capability High reliability Low latency/packet loss between clients and server Used to prevent cheating amongst clients Most commonly used architecture today 13

14. Centralized Architecture 14 Serve r

15. Server responsibilities Authentication Updating positions Maintaining scores/information about players and teams Managing forming of teams 15

16. Online Games Games and Measurement Properties Game System Architecture CentralizedHybrid Networked Games Measurement Challenges State of the Art 16

17. Decentralized architecture 1 Clients interact with each other directly Proposed decentralized architectures:  MiMaze  Mercury  P2P-Support  Zoned Federations 17

18. Decentralized architecture 2 Partial decentralization Partitioning players and associated responsibility into regions Complete decentralization Any peer in P2P network can carry out authentication requirements to eliminate cheating 18

19. Decentralized Architecture 19

20. Online Games Games and Measurement Properties Game System Architecture CentralizedDecentralized Networked Games Measurement Challenges State of the Art 20

21. Hybrid Architecture One example: Mirrored server Each game has several distributed servers Clients only communicate with one of these FreeMMG 21

22. Hybrid Architecture 22

23. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 23

24. Scalability Number of players that can simultaneously participate in a networked game Typical numbers  <10 for RTS (Real Time Strategy)  10-30 FPS (First Person Shooter)  Thousands in MMOGs (Massively Multiplayer Online Games) Increased users  increased delays 24

25. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 25

26. Real-time requirements Often the limiting factor in viability of a game Varying requirements for latency and packet loss Even within a single networked game, different objects may require different real-time standards  e.g., high accuracy sniper rifle vs. machine gun 26

27. Lightening Gun in Unreal Tournament 27

28. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 28

29. Manner of access Wired and Mobile Environment Physical location of client can be used  Increasing popularity of GPS and Bluetooth devices  Require accurate client location abilities  Active Bat, Cricket (indoor location systems), Human Pacman Most games require wired environment for lower latency/packet loss 29

30. Human Pacman 30

31. Game Measurement Properties Measured propertyWhy measuredWhere measured Traffic characterization Growth patterns, popularity Across internet Game system architecture Differences in architecture Across internet Scalability Provisioning, performance Varies with game genre Real-time requirements Game viability and latency limitations Game client Manner of accessMobility constraint Client and server locations Session durationState maintenanceserver 31

32. Single session vs. Multi-session Single session  User connects, plays, then exits game  More common among older games Multi-session gaming  User logs in, plays, stalls session until next game  Value of character can grow or ebb during time  Increases necessity for network performance 32

33. World of Warcraft 33

34. Online Games Games and Measurement Properties Hidden Data Hidden Layers Tools State of the Art 34

35. N etworked Games Measurement Challenges Measurement challenges are significantly different from DNS or Web  High interactivity, lower tolerance for errors and delays Harder to simulate user traffic via programs  Diversity of possible user action Less datasets 35

36. Online Games Games and Measurement Properties Hidden Layers Tools State of the Art 36

37. Hidden Data Skill levels of users  impacts importance of latency/packet loss/etc.  No uniform way to measure impact of network problems Information about game server rarely public, difficult to reverse engineer (e.g. Algorithm server use to arbitrate jobs) Downloading of new content can effect performance Fortunately, usually no intermediaries between client and server 37

38. Online Games Games and Measurement Properties Hidden DataTools State of the Art 38

39. Hidden Layers 1 Games typically involve authentication, setting up parameters, playing, and quitting  One or more steps may be avoided by suspending the state and continuing from a previous session Authentication generally done via TCP handshake Game actions usually sent over UDP or TCP Downloads of game updates sent over TCP Less complex than short session applications (e.g., Web) 39

40. Hidden Layers 2 Quality of game effected by: Network, Client, Server, I/O Devices Delays cause different users to react differently Delays on server end factored into measuring delays from player's view Team games add more complexity to measurements Time of game effects impact of adverse network conditions Location of player changes effect of network problems 40

41. Online Games Games and Measurement Properties Hidden Data State of the Art 41

42. Tools 1 Ping used to measure latency, latency radius (number of active players within latency threshold) Tcpdump for passive recording of game traffic Geographic mapping tools used to locate game servers RTT measured at time of special events such as a player dying 42

43. Tools 2 Traditional passive measurement Average bandwidth Packet inter-arrival time Packet count and size Number of attempted/successful connections Unique clients Non-traditional measurement tools GameSpy tool used to report number of players associated with game server Qstat which display game server status 43

44. Online Games Games and Measurement Properties Networked Games Measurement Challenges Architectural Issues Traffic Characterizatio n Impact of Network Effects Synthetic Traffic Models Generated Mobile Game Environments 44

45. Online Games Games and Measurement Properties Networked Games Measurement Challenges Traffic Characterizatio n Impact of Network Effects Synthetic Traffic Models Generated Mobile Game Environments 45

46. Architecture Issues Decentralized Game Architecture MiMaze Game: Decentralized server research IP Multicast used for player moves Latency limited to 100ms Not a popular architecture  Presence of cheating  Lack of a synchronous view by all players 46

47. Online Games Games and Measurement Properties Networked Games Measurement Challenges Architectural Issues Impact of Network Effects Synthetic Traffic Models Generated Mobile Game Environments 47

48. Traffic Characterization Some Characteristics / Quake and UT Examined characteristics  Inter-arrival time of packets  Size of packets  Differences between client and server were examined Data gathered passively using DAG cards (packet capturing hardware) Client packets were more numerous but smaller than server packets in Quake 48

49. Traffic Characterization Resource Requirements of Game Server / Counter Strike Half a billion packets captured in 1 week from ~6000 players Showed that updates must be predictable to compensate lag Client/server packets maintained properties from Quake study Regular traffic bursts are quite likely Packets size vary Active clients sent relatively uniform load 49

50. Traffic Characterization Player behaviour / Half-Life and Quake Player behaviour studied across a few thousand servers  How many players  How long they connected Time-of-day effects game traffic Players joined games with higher numbers of players Duration a player continue to play appears to be independent of number of players Inter-arrival times of player follows a heavy-tailed 50

51. Time-of-day effects game traffic 51

52. Traffic Characterization Session time / Counter Strike Traces of one-week were gathered with GameSpy  Number of players  Session  Duration Unlike most applications, session times followed a Weibull distribution Most players played for short durations Testing new features of game Study showed difficulty of generalizing network games 52

53. Session Time/Probability Chart 53

54. Online Games Games and Measurement Properties Networked Games Measurement Challenges Architectural Issues Traffic Characterizatio n Synthetic Traffic Models Generated Mobile Game Environments 54

55. Impact of Network Effects Network Latency / Quake 3 Used servers in California and London Intentionally masked London server as California location Found players chose servers closer to them geographically Bottleneck last mile between user and ISP 55

56. Impact of Network Effects Unreal Tournament Emulating packet loss and latency according to live server data Found no significant difference in ability to move due to packet loss (prediction compensation) Even 100ms latency caused drop in perceived performance 56

57. Online Games Games and Measurement Properties Networked Games Measurement Challenges Architectural Issues Traffic Characterizatio n Impact of Network Effects Mobile Game Environments 57

58. Synthetic Traffic Models Generated Each game must be examined and synthesized separately Representative set of players must be found and data captured over a period of time  Skill of players will effect data Typical information gathered  number of packets  packet length  Inter-arrival time  server response time 58

59. Online Games Games and Measurement Properties Networked Games Measurement Challenges Architectural Issues Traffic Characterizatio n Impact of Network Effects Synthetic Traffic Models Generated 59

60. Mobile Environment Human Pacman, Can You See Me Now? Few measurements so far Study on GAV (GPL Arcade Volleyball) game ported to PDA found that wireless environment could not support real time requirements of GAV 60