Networking for Games

INTRODUCTION
Gaming always have been one of
favorite areas for computer scientist, researchers
and developers. Gaming once limited to video
games over console has found new dimensions
and demographics after targeting new trends
such as social network.
Today e.gaming is not limited to
console games. Categories are consoles, PC, 2D,
3D mobile and games over social network are
very famous. Multiplayer games over social
media or network is the latest trend. Social
networking era allow us to involve our friends in
the gaming activities. There are various domain
but games over network introduce a lot of
questions. It arises the question of how data
packets for games will be treated over networks.
Priority to be given on timely delivery of the
packets or in order packet. How the tradeoff
between different priorities and requirements
can be assured without affecting performance
and entertainment.
Its no more just an entertainment.
Gaming has huge share in the software industry.
Gamers are individuals considered as genius
who have good set of skills for solving different
problems, relatively better attention span and
observation skills. Under this paper discussion
will be carried out how networking for games is
handled and what are the aspects of
communication to be considered. Traffic wont
be very different but how packets should be
transmitted and what conditions would be
applied for different genre of games.

PROBLEM STATEMENT:
E.gaming was once considered
notorious; time changed and now science has
accepted the importance of gaming. Today we
have a huge industry invested over gaming.
Along different communication perceptive
problems it arises was data packets traversal as
well packet handling difference between normal
and game traffic. Networking for games and
protocols
add-ons
for
gaming
traffic
implementation along quality of service aspects

in view of different research conducted by

researchers is main focus.
Robust traffic with simple architecture has
always been the goal of networking architects
but this is an ideal goal. Compromising on ether
of them is usually the trade of. Traffic of games
is not very different from the traffic of internet.
Design decisions for multiplayers and online
games are driven from the architecture of
internet. Three main aspects are
Best Effort Service
Addressing
Transport Protocol
These three aspects are no different for
networking for games as well. Approach is to
handle the traffic based on these three aspects
and prioritizing the requirements while using
options available for these three aspects.

LITERATURE REVIEW:
QoS REQUIREMENTS AND CONTENT
BASED TRANSPORT PROTOCOL:
Quality of service requirements for normal
internet traffic and gaming traffic is different.
Authors in [1] analyzed how transport protocols
perform on the action logs of Angels love.
Protocols were UDP, TCP, SCTP, DCCP and
content-based transport protocol.
Gaming protocols have a design goal to
achieve timely delivery along jitter less traffic
with maximum interactivity for better user
experience. Massive multiplayer online role
playing games end to end delay less traffic is
dependent overt transport protocols. TCP and
UDP are widely used for gaming traffic
transport protocol where reliable traffic is not a
requirement UDP is preferred.
Researchers conducted an experiment
on the sample data. They have classified
message types in to main categories and checked
class of message that can tolerate the reliability,
in order and timely delivery failures. The

defined three strategies to overcome and handle

the messages delivery.
Strategy MRO: It uses multi streaming
approach which puts messages in the category
wise stream.
Strategy MR: It uses multi streaming approach
along optional ordering which puts messages in
ordered which are supposed to be delivered in
order else unordered stream
Strategy M: It uses multi streaming approach
along optional ordering and optional reliability.
This handles messages if ordered and reliability
have to be ensure add to ordered and reliable
stream else unordered and unreliable stream.
Three different content-based protocols were
developed on the bases of this strategy and data
log were simulated over existing transport
protocol and the protocols designed on the basis
of the strategies results shown that there were
less jitter and less delay using these strategies.
The reason is strategies are built on the basis on
contents of message and designed specially to
handle the traffic on the bases of content.
Though studies and experiment shown that UDP
and DCCP should be preferred since they gave
better results but TCP is still widely used and
preferred due to its reliable communication.
Content based strategies shown comparatively
less performance to UDP still they maintain a
good game semantics which is key of gaming so
they can be employed in real life models.

OBSCURING

NETWORK

DELAYS IN ONLINE GAMES

AND

JUST

NOTICEABLE

DIFFERENCE:
In multiplayer games with high
interactivity where perspective action of
multiplayers is required its important to
maintain the traffic of network in such a way
that its hardly noticeable by the players and
introduce optimization of actions in order to
ensure real time effect and concealing the
delays.

Gamming has moved from local

networks to the internet. Highly interactive
games are yet to be synchronized since they
have two main characteristics.

Fast speed
Instant
and
requirement.

speedy

judgment

These requirements couldnt afford lag and

highly delay sensitive. Networks characteristics
required to handle these games has to be
designed carefully.
Just Noticeable Difference is the approach to
minimize the effects of delay for multiplayers
and use this approach in online games. Research
was first of its kind since JND has been used in
voice over IP and other kind of traffic but never
used
to
achieve
synchronization
and
optimization for FTG and responsive online
games. Optimizing blank periods and concealing
delay is the major task while designing a
communication protocol or system for online
games.
There are three schemes that can be used for
minimizing and making traffic delay less.
Local Perception filter 1: This schemes
increases the action time of A when the subjects
participated in experiments compared the zero
latency games and with local perception filter
they easily recognized the delay so this schemes
doesnt work so well.
Local Perception Filter 2: In this scheme
response time of the action reception minimized.
If Local Perception filter 1 is applied in
conjunction with LPF1 then this scheme work
well else its very obvious and easily
distinguished
on
purpose
optimization
technique.
Local Lag: The scheme is poor for concealing
delays in fast-moving games.
These schemes were previously used
new proposed scheme in the experiment is JND
where we use these three schemes in different
conjunction. Approach is to bring the awareness
on both side of the player should be equal,
minimized and optimized. Basic Idea is to
enhance quality of experience. Gaming is all

about experience for the user. JND approach is

used to enhance this experience.

Aspects

of

Online

Multiplayer

Games
Multiplayer online games (MOG) are real time
interactive distributed applications. MOG
requires a lot of computational power along with
a strong network.
Mainly MOGs performance is based on 2 things:
Physical platform
Logical platform
Physical platform are the physical components
on which MOG is being played such as
hardware (processor) and networking.
Logical platform on the other hand is the system
architecture that how these hardware
components are organized and how software is
running on them.

processing of game. So even if the network is

strong enough but the computational power is
weak then still game will face a lot of problems.

Logical Platform:
Generally 3 architectures are used for MOGs
Client/Server
Peer-to-Peer
Mirrored/Networked Server
In [4] they have analyzed these system
architectures and have proposed a new
architecture i.e. server-clustered architecture
combining the advantages of the above 3
mentioned architectures.

Physical Platform:
1. Network Resources:
MOGs performance depends on the following 2
major aspects of Network resources
Bandwidth
Latency
Bandwidth requirement mainly depends on the
no of users playing and the transmission
technique like whether packet is being
broadcasted or unicasted. So to improve the
performance multicast is being used to send the
packets only to the intended receivers.
Latency is the delay that is caused due to the
sending/receiving packets from other side of the
system. As a rule of thumb latency between 0.1
and 1 second is acceptable. First person shooters
games requires the minimum latency.
2. Hardware Resources
The underlying systems computational power is
one of the major aspect in affecting the
performance of MOGs.
Computational Power is required to handle 2
things. First to process the incoming network
traffic, second to handle the client side

Clustered Server Architecture:

It replaces the single central server by a group of
connected servers, using this technique the
system becomes more scalable.
In contrast to all of the previous mentioned
architecture where the server has the whole
game world inside it, here each server has partial
game world. Distribution of game world among
these servers here is region oriented. And when
an object crosses a region a hand-off operation
takes place. For the objects which keeps on
changing the boundaries too often then hand-off
operation takes too much resources, so to
minimize this common regions are introduced
between the regions.

Then to speed up the server processing

following aspects were discussed.

Key Issues in Game Server:

1. Interest Management
MOG world consists of thousands of users and
not every user is concerned about every other
user, users at any specific time are playing with
a few no. of players due to which the packets a
player must receive should be from those
concerned players.
To do Interest management most of the servers
use Aura based IM it usually correlates with
the sensing capability of the player.
2. Synchronization
Synchronization is the foundation of MOGs. In
[4] the authors describes the different algorithms
for the synchronization:
Conservative algorithms
Optimistic algorithms
So
by
analyzing
these
algorithms,
synchronization technique is proposed in which
the calculation of each partial world is done at
each server independently and then they send
those to others.

Cheating in MOGs:
Most commonly 3 types of packet tampering is
done in First Person Shooter online games.
Reflex augmentation
Packet interception
Replay attack
Reflex Augmentation: In this when cheater
fires, then the proxy sends some additional
rotation and movement control packets to
improve the aim.
Packet Interception: In this when someone
fires at the cheater, then the proxy prevents those
damage packets from reaching the cheater thus
the cheater feels no damage.
Replay Attack: In this when cheater fires, then
proxy can send that packet for more than 1 time,
increasing the fire rate and thus cheater gets
more chances of hitting the target.

Virtual Team Interactions &

Dynamic Voice Communication
Support
Case: Counter-Strike Multiplayer 3D Action Game
Virtual Environments (VEs) allow simulation of
real-world scenarios and also allow human
interaction with said environments via computer.
When multiple users interact with the
environment simultaneously, then this becomes
a Networked Virtual Environment (Net-VE),
wherein people form virtual teams connected
through a network, regardless of their
geographic locations. These virtual teams are the
focus of this paper, mainly team interaction and
how this virtual interaction emulates real-life
interactions. To carry out this research, the
author has chosen a case study of a virtual
environment-based game, Counter Strike,
which is a multi-player 3D game that preserves
spatio-temporal aspects to give it a real-life look
and feel. This provides a way for the virtual
team to interact in the context of the game but
also limits the ways they can interact as part of
the game.
Author has proposed a taxonomy on the
interaction of teams within virtual environments,
which includes the following: a shared sense of
space, where despite being separated
geographically, the team feels connected as if
they share space; a shared sense of presence,
which is ensured through avatars of every
member of the team; a shared sense of time,
which allows real-time interaction; a way to
communicate, which allows users to interact
without interacting in real life; a way to share,
which the author describes as the dynamic
environment that allows interaction. After
studying multiple virtual environment-based
games, the author was able to generalize the
categories of interaction as follows: Avatar
presence, Control and co-ordination, objectbased interactions, autonomous and automatic
actions, gestures, physical contact, body

language, language-based communication and

world modifications.
This paper uses qualitative research, wherein the
empirical data was gathered by interviewing
Counter Strike participants during a LAN
gaming event. The target group of players were
over the age of 16 and they were organized in a
way that they couldnt see all of their team
mates. Counter Strike (CS) is a rather realistic
combat simulation game, where multiple
players work in teams to achieve their goal,
which may vary, like defusing bombs or
rescuing hostages etc. The observed team
interactions showed that for this game, avatar
presence is a graphical representation of team
mates, and also shows the role the member
plays. Also, posture changes occur in avatars
that let players know what other players are
doing, and may warn them to do the same, for
instance hiding behind a crate could alert others
to take cover. Others players can also find out
the weapons through the visible outlook of the
avatars. Autonomous & Automatic Actions is
also an important feature in the game, where the
avatar shows the status of the player through
actions, eg. a limp in case a player has an injured
leg. Self-programmed actions (cheat codes) are
frowned upon in this community. Body
Language is not very important as the game is so
fast-paced, players do not spare a lot of time on
gestures.
Control and co-ordination plays a major role in
team interactions, since the game is very teamoriented. The co-ordination, however, is left to
the players themselves and is not forced upon
them by the game. The players co-ordinate
amongst themselves voluntarily to achieve more
favorable results. An example of this would be
creating a diversion when one or more team
members are in trouble. Gestures, according to
research, is one of the most important factors of
such games, but CS doesnt support it very well.
However, players believe it will be extremely
useful, like pointing the direction of the weapon.
Language-based communication is restricted to
only pre-recorded messages, like getting help for
the wounded. And any additional interaction

between players can be through textual chatting.

The lack of this feature has been overcome using
voice-over-IP
messaging.
Object-based
interactions, as in the handling of weapons and
the player interaction with the weapons they
selected at the beginning of the game brings the
element of role-playing in the game. Selecting,
reloading and using bombs are examples of
object interaction. World modification is not a
part of this game, even though in some cases,
doors/windows can be destroyed. Seeing as the
game is quite aggressive, and even violent,
physical contact is a large part of this game,
where main examples of contact are shooting
and stabbing. Non-violent physical contact is
rare, in the form of touching hostages, etc.
The research found that counter-strike is
definitely a team-oriented game, which allows
many ways of team interaction. Teams
communicate and co-ordinate themselves and
have their own self-assigned roles and strategies.
And since communication is a little limited, it
was observed that teams generally invent unique
ways of communicating among themselves. The
general consensus among players was that they
love the satisfaction of beating an opponent and
have fun actually discussing their tactics after
the game is over. An interesting research finding
was that when there is more at stake during the
game, as in, if in a tournament, interaction
among players increases. As far as verbal
communication is concerned, some players like
the game the way it is, while others would like
to see verbal communication integrated into the
software. However, aside from lack of verbal
communication, non-verbal communication is
also not supported in the game.

CONCLUSION:
Multiplayer online games are the latest trend
nowadays in the gaming world. But very limited
research has been done on it, especially on the
networking of games. Therefore networking for
games should be highly focused to contribute in
the revenue of massive industry and enhancing
players experience.

REFERENCES:
[1] Chih-Ming Chen, Te-Yuang Huang, Kuan-Ta
Chen and Polly Huang, "Quantifying the Effect
of Content-based Transport Strategies for Online
Role Playing Games," In Proceedings of ACM
NetGames 2008 (Poster), 2008.
[2] XU, Jingxi, WAH Wan Sang Benjamin.
Concealing Network Delays in Delay-Sensitive
Online Interactive Games Based on JustNoticeable Differences Paper presented in the
IEEE ICME, International Conference on
Multimedia & Expo, 2013, 6 pgs
[3] Jouni Smed, Timo Kaukoranta, Harri
Hakonen,
(2002)
"Aspects
of
networking in multiplayer computer
games", The Electronic Library, Vol. 20
Iss: 2, pp.87 97
[4] On the Design of Multiplayer
Online Video Game Systems, Chia-

chun Hsu, Jim Ling, Qing Li and C.-C.

Jay Kuo, Integrated Media Systems
Center and Department of Electrical
Engineering
University of Southern California, Los
Angeles, CA 90089[5] Tomas Hildebrandt, Sonja Bergstrer,
Christoph
Rensing,
and
Ralf
Steinmetz. Dynamic
voice
communication
support for multiplayer online games.
In Proceedings of ACM NetGames
2008, pages 102103. ACM Press, 2008.
[6] Tony Manninen. Virtual Team Interactions in
Networked Multimedia Games Case:
''Counter-Strike'' Multi-player 3D Action
Game. In Proceedings of PRESENCE2001
Conference, 2001.