Aspects of Networking in Multiplayer Computer Games

1
Aspects of Networking in Multiplayer Computer
Games

• Jouni Smed, Timo Kaukoranta, Harri Hakonen
• Presented by
• Sanghyun Yoo
• Database Laboratory

2
Contents

• Introduction
• Networking Resources
• Distribution Concepts
• Scalability
• Security and Cheating
• Summary

3
Introduction

• Distributed, real-time multiplayer computer games
  (MCGs)
• in the vanguard of utilizing the networking
  possibilities
• related research have been done
• in military simulations, virtual reality systems,
  and computer supported cooperative working
• the suggested solutions diverge from the problems
  posed by MCGs

4
Networking Resources

• Bandwidth
• the proportion of the amount of data transmitted
  or received per unit time
• bandwidth requirements depend on
• frequency and size of messages
• the number and distribution of users
• the transmission technique (unicasting,
  multicasting, broadcasting)
• Latency
• the length of time (or delay) that incurs when a
  message gets from one designated node to another
• cannot be totally eliminated
• Computational Power
• for handling network traffic

5
Distribution Concepts (1/6)

• Communication Architectures
• single node
• peer-to-peer

• client/server
• server-network

c
6
Distribution Concepts (2/6)

• Data and Control Architectures
• centralized architecture
• only one node holds the data
• must use a two-way relay for networking due to
  the consistency requirements
• distributed architecture
• each node holds a subset of the data
• player controlled entities, whose behavior is
  unpredictable and for whom there can be only one
  source of commands
• indeterminism

7
Distribution Concepts (3/6)

• Data and Control Architectures (cont'd)
• replicated architecture
• a copy of the same data exist in all nodes
• non-player characters and other computer
  generated entities are predictable and need not
  send frequent control messages
• determinism
• both of distributed and replicated architecture
• use the short-circuit relay and provide higher
  responsiveness

8
Distribution Concepts (4/6)

• Compensatory Techniques
• Message Compression and Aggregation
• saves bandwidth
• requires extra computation and weakens the
  responsiveness
• Interest Management
• allows the nodes to express interest in only the
  subset of information that is relevant to them
• aims at reducing the number of transmitted
  messages by specifying the interested receivers
• aura
• the area of interest
• an subspace where interaction occurs
• can be divided further into a focus and a nimbus

9
Distribution Concepts (5/6)

• Focus and Nimbus

• Auras (areas of interest)

10
Distribution Concepts (6/6)

• Compensatory Techniques (cont'd)
• Dead Reckoning
• predicting the data from the other nodes
• which allows to prolong the interval of message
  transmissions and abolish the network latency at
  the cost of data consistency
• the messages can be sent only when dead reckoning
  exceeds some error threshold

• Dead Reckoning

11
Scalability (1/3)

• Serial and Parallel Execution

• No Serializable Part

• No Parallel Part

• Parellel and Serial Computation

12
Scalability (2/3)

• Serial and Parallel Execution (cont'd)
• for serializable parts
• the nodes must agree on the sequence of events
• the communication capacity in a client/server
  architecture using unicast
• d f n C
• d the number of bits in a message
• f the transmission frequency
• n the number of unicast connections
• C the maximum capacity of the communication
  channel

13
Scalability (3/3)

• Communication Capacity

m the number of servers, n the number of clients
14
Security and Cheating (1/2)

• Packet and Traffic Tampering
• reflex augmentation
• the proxy replaces human reaction to produce
  superior results
• e.g., aiming proxy
• packet interception
• the proxy prevents certain packets from reaching
  the cheating player
• e.g., suppressing the packets containing damage
  information
• packet replay
• the same packet is sent repeatedly
• e.g., the fire command packet

15
Security and Cheating (2/2)

• Information Exposure
• to utilize the server
• which can check whether a client issuing a
  command is actually aware of the object with
  which it is operating
• democratized version
• Design Defects
• can create loopholes which the cheaters are apt
  to exploit

16
Summary

• Aspects of MCGs
• networking resources
• bandwidth, latency, computational power
• distribution concepts
• communication, data and control architectures
• compensatory techniques
• scalability
• parallel and serial parts
• communication capacity
• the problems of ensuring security in MCGs
• tampering, information exposure, design defects

17
On the Impact of Delay on Real-Time Multiplayer
Games

• Lothar Pantel, Lars C. Wolf
• Presented by
• Sanghyun Yoo
• Database Laboratory

18
Contents

• Introduction
• Multiplayer Games
• Experimental Evaluation
• Summary

19
Introduction

• Multiplayer Games
• several persons interact simultaneously over
  networks like the Internet
• network transmission delay
• a major problem of networked multiplayer games
• cannot be avoided completely
• it is important to investigate the impact such
  delay can have on the performance of multiplayer
  games

20
Multiplayer Games (1/2)

• Delay
• the time between the generation of an event and
  the resulting update of the game state including
  the presentation to the users
• leads to significant problems for real-time
  multiplayer games

21
Multiplayer Games (2/2)

• Experiments with Commercial Games
• Re-Volt and Need-for-Speed
• delay of approximately 100ms and 200ms
• five tests
• start release
• start process
• simultaneous starting
• driving with constant speed
• collision treatment
• observed results
• the games provided very limited handling of delay
  effects only

22
Experimental Evaluation (1/9)

• Principle Setup
• to enhance the consistency
• delay the presentation of the actual game state

23
Experimental Evaluation (2/9)

• Implementation

24
Experimental Evaluation (3/9)

• Performed Experiments
• four aspects
• average time per round in dependence on delay
• best time per round in dependence on delay
• driven course
• frequency of leaving the course

25
Experimental Evaluation (4/9)

• Results
• average time per round

26
Experimental Evaluation (5/9)

• Results (cont'd)
• best times per round

27
Experimental Evaluation (6/9)

• Results (cont'd)
• mean of average round time

28
Experimental Evaluation (7/9)

• Results (cont'd)
• frequency of departure from course

29
Experimental Evaluation (8/9)

• Results (cont'd)
• driven course
• drivers who had some rolling motions
• entered a zigzag course from 100 ms to 250 ms
  delay
• drivers who avoided rolling motions
• entered a zigzag course from 100 ms to 300 ms
  delay

30
Experimental Evaluation (9/9)

• Results (cont'd)
• subjective impression

31
Summary

• The Impact of Delay
• a delay up to 50 ms is uncritical for a
  car-racing game
• by the objective measurements
• by the spontaneous statements of the participants
• a delay of more than 100 ms should be avoided
• in the overview view
• if the system should provide for some realistic
  driving behavior