Clan Web Sites, Online Games and Hosting

Broadband Access Networks Network capacity is not always the same in both directions. The oldest access network application is telephony. With telephony, capacity is usually the same in both directions. Parties involved in a conversation talk, approximately, the same amount during a conversation. Of course, this is not always the case, and some conversations may be quite one-sided. Nevertheless, telephone systems have evolved with the assumption that capacity in both directions needs to be the same. However, many new applications, including most Internet games, require more capacity in the network to the user direction (the forward link) than in the user to the network direction (the reverse link). Typically, a player will receive a great deal of information about the state of the game from the game server on the forward link but will transmit relatively little information (mainly player actions) on the reverse link. Some access networks such as ADSL and cable modem networks naturally support this asymmetry. Their forward link capacity is much greater than their reverse link capacity. Other technologies, such as some cellular broadband networks, are capable of adapting to this asymmetry while other networks provide only symmetric capacity. Very few access networks are able to offer Quality of Service guarantees. That is, the network guarantees that important connection characteristics such as bit rate, delay and delay variation will be within agreed limits. Specification of these limits form a spectrum ranging from best effort networks where there are no guarantees through networks where guarantees are specified in terms of long-term averages, to networks where bit rates, error rates, delays and variation in delays are tightly specified. As we move across this spectrum, networks generally become less commonly available and more expensive to implement and to use. Access networks are also often a major source of latency and jitter. Latency is the delay experienced in a game before a command by the user is reflected in the game state. Jitter is the variation in delay. Networks where the same medium is shared among many users, such as wireless and cable modem networks, are more prone to variations in delay than other networks. Often users will want to access a game while they are mobile. Perhaps they are travelling on public transport, or they might be passengers in a car. Some access networks are able to support mobile users more effectively than others. Although there have been many interesting developments in Mobile Internet Protocol (IP) networks, cellular wireless networks are currently most effective in supporting mobile users. Some access networks are more convenient to access and to use than others. For example, connecting to a cellular network is usually trivial. The telecommunications company running the network provides direct access to the Internet. However, connecting through a Wireless Local Area Network (WLAN) may require significant reconfiguring of the device used to connect to the network. In this chapter, we provide an outline of the strengths and weaknesses of the most commonly used access network technologies. We discuss their design, capacity, potential to introduce latency and jitter, reliability, usage costs and how they are typically used. We hope that through understanding these networks, game developers and those deploying games will have a better understanding of the network constraints within which their games must operate. 8.2 Access Network Protocols and Standards In this section, we discuss some of the general characteristics of access networks.

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Inexpensive Web Hosting services

124 Networking and Online Games: Understanding and Engineering Multiplayer Internet Games Access networks are concerned with the delivery of framed packets from the user s network connected device to the network. Access networks are differentiated from each other by their physical and data link layers. 8.2.1 Physical Layer A network s physical layer specifies the transmission medium and how individual bits are encoded onto it. The media used in access networks are wireless, coaxial cable, twisted copper pairs and optic fibre. The physical layer may also be responsible for minimising or even correcting errors as well as controlling access by multiple users. Physical media include wireless where communications are mapped onto Radio Frequency electromagnetic radiation. For example, WLANs use a range of frequencies denoted as the Industrial, Scientific and Medical (ISM) Bands of around 2.4 GHz or 5.8 GHz. Cellular networks use frequencies around 800 to 900 MHz or 1900 MHz, depending on the particular network. Coaxial cable is another commonly used medium. This is most commonly used in cable modems where the access network is overlayed on cable television network infrastructure. Another broadband access network overlaid on an existing network is ADSL, which uses the twisted pair last hop of the Public Switched Telephony Systems Network (PSTN). Using sophisticated coding and error correction techniques, high bit rates can be obtained over the existing twisted pair medium. The other commonly used medium is fibre optic cable. With this technology, the medium comprises very thin lengths of glass or plastic down which high-speed lasers or light-emitting diodes send light as the signal. Fibre optic cable is often used in hybrid fibre-coax Cable TV networks. The physical layer specifies how individual bits are mapped onto the physical medium. The process of mapping a bit or a series of bits onto a medium is referred to as modulation. Wireless systems may encode bits as variations in the amplitude, phase or frequency of a carrier wave. Similar modulation schemes are used for coaxial cable and other electrical wire media. Bits transmitted across a fibre optic cable are typically encoded as pulses of light. The best method of modulation depends on the characteristics of the media it will use and the purpose of the network. The physical layer specifies how communication channels are mapped onto the physical medium. There needs to be some way of dividing the physical medium into multiple channels. The process of mapping multiple channels onto one physical medium is called multiplexing. Multiplexing can be done using frequency division, time division, code division or a combination of all three. In frequency division multiplexing (FDM), communications are modulated onto a carrier wave of a specified frequency and transmitted. The receiver filters out all frequencies other than that of the specified carrier wave, then demodulates the carrier wave to recover the original signal. Time division multiplexing (TDM) involves allocating the medium to each channel for short periods of time. These periods of time are commonly called timeslots. TDM can be synchronous, where each channel is allocated a fixed number of timeslots in a given period, or asynchronous, where each channel is allocated a variable number of timeslots in a given period.

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Inexpensive Web Hosting services

122 Networking and Online Games: Understanding and Engineering Multiplayer Internet Games Game Client Server Game client ISP ISP server router Access network The Internet router Access network Figure 8.1 Access network between different nodes in the network, which may be located vast distances apart. Access networks take comparatively smaller traffic volumes from geographically distributed users and deliver it in aggregate form to the core network. There is no one best solution to connecting users to a core network. The most appropriate solution will depend on many different factors, including the population density of users to be connected, the distance to the nearest core network node (such as a telephone exchange or wireless base station), and existing installed network infrastructure amongst other factors [ARM2000]. Consequently, there is no single broadband access network. Instead, there are a number of different network technologies that come under the general category of being access networks. Not all kinds of access networks are available in all places at all times. For example, coverage of some broadband wireless networks is often limited. Wireless local area network (LAN) for example, has a range of only tens of metres. Availability of some networks depends on what other infrastructure has been deployed. For example, cable modem networks have usually been deployed in the wake of cable television deployment. Asymmetric Digital Subscriber Link (ADSL) networks are usually deployed on existing two-wire telephony infrastructure. A game developer cannot assume that their player market will either use, or have available for use, a specific access network. Not all access networks have the same level of reliability. Wireless transmission, although having many advantages over wired transmission, is inherently unreliable [RAPP2002]. Generally, this unreliability is controlled through transmitter and receiver design for constructing a reliable wireless network that provides a satisfactory service to customers, but nevertheless, game players are likely to experience a greater level of unpredictable variations in delay and capacity if their access network is wireless-based than if it is cable based or fibre optic based. Usage costs of different broadband networks vary tremendously. The infrastructure needed, the number of users supported by the network and competition from other technologies and service providers, all affect prices that individual users will be charged. Usage costs themselves can be charged in different ways time-based, volume based, a regular flat fee, or a combination of all three. Although it is difficult to generalise, broadband cellular networks, because of the vast infrastructure investment needed to provide global coverage and seamless mobility support, are typically much more expensive to use than other networks. Different networks have different capacities. Generally, cellular wireless networks have much lower bit rates than wired networks. Also, some broadband technologies use a shared medium, which means that, although total capacity might be constant, the capacity available to an individual user depends on the number of other users.

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Clan Web Hosting services

Playability versus Network Conditions and Cheats [CHACK] Counter Hack, http://www.counter-hack.net/, Accessed 2006. [CHO2001] C Choo, Understanding Cheating in Counterstrike , (online) http://www.fragnetics.com/articles/ cscheat/print.html, November 2001. [DUMMYNET] Dummynet Traffic Shaper, Bandwidth Manager and Delay Emulator , http://www. FreeBSD.org/cgi/man.cgi?query=dummynet&sektion=4, Accessed 2006. [FBKERN] Chapter 8 Configuring the FreeBSD Kernel, FreeBSD Handbook, 2005 (http://www.freebsd.org/ doc/en US.ISO8859-1/books/handbook). [HEN2001] T. Henderson, Latency and user behaviour on a multiplayer game server , Proceedings of the 3rd International Workshop on Networked Group Communications (NGC), London, UK, November 2001. [HEN2003] T. Henderson, S. Bhati, Networked games a QoS-Sensitive Application for QoS-insensitive Users? ACM SIGCOMM RIPQoS Workshop 2003, Karlsruhe, Germany, August 2003. [IPTABLES] The Netfilter/iptables Project , http://netfilter.samba.org, Accessed 2006. [MADDEN] Madden NFL 2004 on EASPORTS.com , http://www.easports.com/games/madden2004/, 2004. [NIC2004] J. Nichols, M. Claypool, The Effects of Latency on Online Madden NFL Football , Proceedings of the 14th ACM International Workshop on Network and Operating Systems Support for Digital Audio and Video (NOSSDAV), Kinsale, County Cork, Ireland, June 16 18, 2004. [NISTNET] National Institute of Standards and Technology, NIST Net Home Page , USA, http://snad.ncsl. nist.gov/itg/nistnet/, Accessed 2005. [OLI2003] M. Oliveira and T. Henderson, What Online Gamers Really Think of the Internet , Proceedings of the 2nd Workshop on Network and System Support for Games (NetGames 2003), Redwood City, CA, USA, May 2003. [PBUSTER] PunkBuster Online Countermeasures , http://www.punkbuster.com/, Accessed 2006. [RIZ1997] L. Rizzo, Dummynet: A Simple Approach to the Evaluation of Network Protocols , ACM Computer Communication Review, Vol. 27, No. 1, pp. 31 41, 1997. [SHE2003] N. Sheldon, E. Girard, S. Borg, M. Claypool, E. Agu, The Effect of Latency on User Performance in Warcraft III , Proceedings of the 2nd Workshop on Network and System Support for Games (Netgames 2003), Redwood City, CA, USA, May 2003. [STEAM] Welcome to Steam , http://steampowered.com/, Accessed 2006. [VAN2003] W.A. Vanhonacker, Evaluation of the FreeBSD Dummynet Network Performance Simulation Tool on a Pentium 4-based Ethernet Bridge , CAIA Technical Report 031202A, December (http://caia.swin.edu.au/ reports/031202A/CAIA-TR-031202A.pdf), 2003. [YAN2002] J. Yan, H-J Choi, Security Issues in Online Games , The Electronic Library: International Journal for the Application of Technology in Information Environments, Vol. 20, No. 2, 2002, Emerald, UK. [YAN2003] J. Yan, Security Design in Online Games , Proceedings of the 19th Annual Computer Security Applications Conference (ACSAC 03), IEEE Computer Society, Las Vegas, USA, December, 2003. [YAN2005] J. Yan and B. Randell, A Systematic Classification of Cheating in Online Games, 4th Workshop on Network & System Support for Games (NetGames 05), Hawthorne, New York, USA, pages 10 11, Oct 2005. [ZAN2004] S. Zander, G. Armitage, Empirically Measuring the QoS Sensitivity of Interactive Online Game Players , Australian Telecommunications Networks & Applications Conference 2004, (ATNAC2004), Sydney, Australia, pages 8 10, December 2004.

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Clan Web Hosting services

Broadband Access Networks In this chapter, we examine the networks used to connect users to the Internet. These networks are commonly known as Access Networks. There is no single network technology that comprises an access network. Access networks can be based on wireless, coaxial cable, existing telephony cables or fibre optic cable. Access networks might be run by telecommunications companies, corporate entities, private users or a combination of all three. There are many different ways in which users can gain access to the Internet. In this chapter, we survey the most important ones. 8.1 What Broadband Access Networks are and why they Matter Our interest in this chapter is in broadband access networks. In this context, broadband refers to the bit rate at which the network connects the user. There is no widely accepted agreement as to what bit rate constitutes broadband access. Some definitions have it as a minimum of 256 kbps, others 512 kbps and still others at 1 Mbps or more. For the purpose of this chapter, we will refer to broadband as being at least 256 kbps. 8.1.1 The Role of Broadband Access Networks Figure 8.1 shows in a simplified form the relationship between the game client, the access network, the Internet and the game server. The client is connected to their Internet Service Provider s (ISP s) router across their access network. Similarly, the game server is connected to their ISP through another access network. It is important that those involved in developing and deploying games have an understanding of the strengths and weaknesses of the many different kinds of access networks. Different access networks have quite different characteristics, which can have impacts on the quality of the game experience. Depending on the access network, game players may experience quite different effects with regard to latency, jitter, capacity and usage costs. 8.1.2 Characteristics of Broadband Access Networks Access networks differ from core networks in a number of ways. Generally, these differences are a consequence of the Last Mile problem the cost of providing connectivity to a geographically distributed population. Core networks shift large volumes of traffic Networking and Online Games: Understanding and Engineering Multiplayer Internet Games Grenville Armitage, Mark Claypool, Philip Branch . 2006 John Wiley & Sons, Ltd

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Clan Web Hosting services

Playability versus Network Conditions and Cheats In principle, a malicious player could attempt to flood through the game itself (e.g. sending in-game text messages to other players at a high rate, which could create the desired flood of traffic at the IP layer). However, a centralised game server can impose limits on the rate at which it forwards in-game messages between players. A reasonable rate limit (on the basis of the ability of the human recipients to read legitimate messages) would be far below the rate needed to flood a victim s underlying network links. There are two broad approaches to mitigating client-side cheats minimise the amount of game-state information proactively distributed by the server, and verify that the client s operating environment is within normal parameters. The following trade-off enables wallhacks to provide tactical advantage: There is information a player should not see until the right time . To ensure a player sees new events as quickly and smoothly as possible, the server s snapshots carry information about objects and players not yet visible to the player. The client software is trusted to only reveal this additional information at an appropriate time and in an appropriate manner. This set of trade-offs are understandable, particularly as they minimise the processing load at the game server and maximise the client s ability to reveal in-game objects the instant they should come into view. One solution would be for game servers to only send to each client snapshots pertaining to in-game entities that each player should be able to see. This would negate the value of most wallhacks as the client will not be told what is behind objects that the server believes are opaque. Unfortunately, there are two problems with this solution. First, this increases the processing required to calculate each snapshot at the game server. Second, the server s knowledge of each player s field of view lags by at least one RTT. Thus, the server cannot send data about other objects coming into view until after the player has, for example, walked around a corner. This would create a somewhat jarring experience (particularly in FPS games), and likely disadvantage players with high RTT relative to those with low RTT. A compromise is for the game server to only include in each snapshot the entities the player can see and entities the player is likely to see within a very short period into the future (given the trajectories of the relevant entities). The trade-off is this: By reducing how far into the future a server s snapshots see you certainly reduce how much information can be revealed by a wallhack; yet you also reduce the ability of the client to smoothly render other entities coming into view. The converse is also true. An alternative approach is to verify that the client s operating environment is within normal parameters. This is the principle behind PunkBuster [PBUSTER] and similar services. A third-party application inspects the client software s integrity, client-side configuration variables and the video software/hardware combination to verify that no known cheats are being employed. The cheat-detection application will usually monitor the client host s operating environment continuously throughout a game, rather like an embedded anti-virus tool. For example, PunkBuster (available for 14 FPS games at the time of writing) has both client and server components that communicate regularly throughout a game to ensure that client-side parameters are within bounds as appropriate for the game. PunkBuster s client and server components also regularly call home to the main

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Mac Web Hosting services

118 Networking and Online Games: Understanding and Engineering Multiplayer Internet Games PunkBuster site for updates and upgrades, thus automatically staying abreast of the latest cheats. Current anti-cheat tools are quite sophisticated. Yet, ultimately they too can be corrupted because, as noted earlier, all client-side software runs on physical hardware entirely under control of the player. At best, tools like PunkBuster can only aim to be so complicated and comprehensive that potential cheaters are discouraged from even risking an attempt due to the chance of being caught. This leads to the question of consequences. Anti-cheat tools are neutered if there are no consequences to being caught. Valve, developers of Half-Life 2 utilise Valve Anti- Cheat ( VAC ), (although the current version is sometimes known as VAC2 ) to detect corruption of client-side environment and explicitly ban players caught cheating. Because of the tight integration of VAC and Valve s Steam player authentication system [STEAM], VAC banning is quite effective. PunkBuster can also keep a record of cheaters, and impose long-term bans if the underlying game s authentication mechanism allows long- term identification of a particular player. Given the feasibility of introducing severe consequences for cheating, another approach to cheat-mitigation is to verify a game s integrity after cheating has occurred. Although this does not prevent someone from cheating in the first instance, the risk of being later discovered and banned from a game should reduce the incentive for people to try cheating. Proposals along these lines have been made for RTS games [CHA2005]. Players might regularly create cryptographically hashed versions of their map view and own movements and share this information with other players during the game. Clear-text information is only sent when other players claim to have viewable areas overlapping your units. Hashing prevents the other players from extracting information about your units that they should not know at a particular point in time. At the end of the game, each player shares their actual moves and the temporary secret key they used to generate their hashed information. The validity of every other player s claimed viewable area at every step of the game can thus be verified through the (now decoded) log of hashed move information. More information about cheats and anti-cheat techniques for a range of games can be found in many online forums (for example, Counter-Hack [CHACK]). References [ARM2001] G. Armitage, Sensitivity of Quake3 Players To Network Latency , Poster session, SIGCOMM Internet Measurement Workshop, San Francisco, November 2001. [ARM2003] G. Armitage, An Experimental Estimation of Latency Sensitivity in Multiplayer Quake 3 , 11th IEEE International Conference on Networks (ICON 2003), Sydney, Australia, September 2003. [ARM2004A] G. Armitage, L. Stewart, Limitations of using Real-World, Public Servers to Estimate Jitter Tolerance of First Person Shooter Games , ACM SIGCHI ACE2004 conference, Singapore, June 2004. [ARM2004B] G. Armitage, P. Branch, L. Stewart, Mathematical and Experimental Analysis of Limitations in Creating Artificial Jitter for Networked Usability Trials , 3rd Workshop on the Internet, Telecommunications, and Signal Processing (WITSP 04), Adelaide, December 20 22, 2004. [BEI2004] T. Beigbeder, R. Coughlan, C. Lusher, J. Plunkett, E. Agu, M. Claypool, The Effects of Loss and Latency on User Performance in Unreal Tournament 2003 , ACM SIGCOMM 2004 workshop Netgames 04: Network and System Support for Games, Portland, USA, August 2004. [CAR2003] M. Carson, D. Santay, NIST Net: A Linux-based network emulation tool , ACM Computer Com munication Review, Vol. 33, No. 2, pp. 111 126, 2003. [CHA2005] C. Chambers, W.C. Feng, W.C. Feng, D. Saha, Mitigating Information Exposure to Cheaters in Real Time Strategy Games , Proceedings of the 15th ACM workshop on network and operating systems support for digital audio and video (NOSSDAV 2005), Skamania, Washington, USA, June 2005.

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Mac Web Hosting services

116 Networking and Online Games: Understanding and Engineering Multiplayer Internet Games penalises disconnections before a game ends). Or the attack could turn the victim into a high ping player, making him or her undesirable to the team on which they had been playing. This problem is not limited to FPS games. It has been noted that … many online Go players choose to play 25 stones in 10 minutes, and it is not unusual for them to play five stones in the last 10 seconds. During the 10 seconds, the [flood attack] is always deadly. Most victims lose their games, and they are not sure whether they are attacked or the network connection is jammed due to other normal reasons. [YAN2002]. Network traffic is relatively easy to intercept. This truth lends itself to proxies that assist in game-play without requiring any modification to one s game client. Aimbots have been implemented as network-layer proxies on this basis intercepting and interpreting snapshot packets in one direction and rewriting shoot command packets in the other direction. Another form of cheating is to connect as multiple users at the same time and collude with each other . For example, one person could play as both partners in a game of online Bridge [YAN2003], thereby having an out-of-band path between the two partners who nominally should have no way to communicate in private. 7.2.5 Cheat-mitigation A central principle of online, multiplayer game design should be: Do not delegate more trust than is absolutely necessary. There will always be someone in a position to abuse trust. The goal is to spread risk by minimising the amount of control different people have, while still maintaining the desired level of interactivity and immersion/realism. (This is actually a good design principle for all network-mediated communication systems.) Some server-side cheats are mitigated by discovering (and fixing) incomplete in-game interpretation of game-play rules. Other cheats can be avoided by deploying appropriate personnel security practices with respect to employees who maintain the servers (e.g. authenticated audit trails on database or configuration changes). At the end of the day, server-side cheats are self-correcting an easily abused server quickly becomes unpopular and either dies through lack of use, or becomes a customer-support nightmare for the hosting company. Network-layer DoS attacks rely on knowledge of the victim s IP address. Anything that hides player IP addresses will reduce the likelihood of such attacks. This is one significant advantage of client server network communication model only the game server needs to know each player s IP address, and each player only needs to know the server s IP address. Games utilising a peer-to-peer communication model at the network layer by definition must expose every player s IP address to every other player. Note that a game may employ a peer-to-peer system model whilst employing a client server network model. The peers can utilise a central server at the IP layer to pass game-state updates between themselves without having to reveal their own IP addresses to each other. (Using IP multicast would not help here. Although each player does not need to know anyone else s IP address to send packets, each player s own IP address is revealed in the source address field of packets they send to other clients.)

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Mac Web Hosting services

Playability versus Network Conditions and Cheats client operates. At the end of the day we can only reduce the likelihood of client-side cheats rather than remove them entirely. 7.2.4 Network-layer Cheats The Internet has a number of characteristics that benefit potential cheaters. There is not a great deal of control over resource consumption both how much you choose to consume, and how much consumption is imposed upon you. There is not much attention to identity, authentication and privacy at the IP layer either. Perhaps one of the most annoying cheats is disruption of another player s network connection by flooding it with excess IP packet traffic. Since many players are connected via dial-up or consumer broadband ADSL or Cablemodem links, even a minimalist application of distributed denial of service (DDoS) techniques can be quite effective. The main goal of a DDoS attack is to overload some part of the victim s network access path, leading the victim to experience a large spike in latency and packet loss rates. Depending on how it is applied, the victim may not even consider the degraded service to be unusual. In most cases the victim has no way of blocking the inbound flood of traffic before it reaches, and saturates, the weakest part of the victim s Internet connection. Figure 7.8 shows how a flooding/DDoS attack could be launched against an unsuspecting victim. In principle, the cheater can launch the attack from anywhere, keeping their own game client s network connection free of excess traffic. Increasing the victim s latency is intended to lead directly to degraded game play (particularly in FPS games) and eventual disconnection of the victim from the game. The flooding attack may be an end unto itself, or simply a means to trigger a disconnection that will by itself cause negative consequences for the victim (e.g. if the game server Victim Game server Cheater Host B Host A Influx of packets floods victim’s access link, causes latency spike and packet loss Internet Flood packets Game traffic Cheater acts alone, or co-opts other hosts to magnify flood traffic Figure7.8DDoSbeingtargetedatavictimwhileplayingagame

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Mac Web Hosting services

Playability versus Network Conditions and Cheats Figure 7.5 Scene from Quake III Arena regular view (top) and glow hack (bottom). Reproduced by permission of Id Software, Inc. Figure 7.6 Close-up of a Quake III Arena character regular view (left), name and weapon revealed (right). Reproduced by permission of Id Software, Inc. Wallhacks can be achieved by modifying game client code directly or modifying external software, such as video device drivers, on which the client relies to render images on screen. Today s graphically advanced games offload much of their actual 3D graphics rendering to the operating system s graphics drivers. Opaque objects are drawn by sending a

Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Mac Web Hosting services