Networks and Protocols

The goal of computer communication networks is to enable efficient transfer of information among a group of people and/or computers. The motivation for building these networks comes from both economical (e.g. sharing of resources, reliability) and social benefits (e.g. increased access to public information, person-to-person communication independent of geography). However, the software and hardware systems which are the building blocks of communications networks are inherently complex. Significant research and engineering effort is required to design, build, test and maintain these systems, to ensure a reliable communications service is available to users.

Background

Types of Networks

Computer communication networks can be classifed in a number of ways, the most common is by geographical size (which also relates to number of users as well as performance characteristics). The table below illustrates a common classification of networks, and lists some example technologies for these types of network. Other classifications include: access networks; transport networks; backbone networks etc.

Network Size Usage Data Rates Example
Personal Area Networks (PANs) several metres Communicate among devices within vicinity of a person kb/s to Mb/s Bluetooth, USB
Local Area Networks (LANs) 10's to 1000 metres Office, home, street communications 10's to 1000 Mb/s Ethernet, Wireless LAN
Metropolitan Area Networks (MANs) km's Inter-office and building connections; city wide telecommunications coverage Gb/s GSM, ATM, SDH/SONET
Wide Area Networks (WANs) km's to global Connections between cities, countries Gb/s to Tb/s Satellite, SDH/SONET

The transmission technologies used in the different networks can include: electrical transmission over copper telephone lines; fibre optics; terrestrial radio; and satellite transmisison.

Protocol Hierarchies

To cope with the inherent complexity, communication networks are often separated into a number of layers, which forms a communications architecture. Each layer provides a service to the next higher layer while hiding the complexities within the layer, and the layers below. Protocols, a set of rules and methods for communication, are defined for each layer. This layered approach offers two main benefits:

  1. Reduced complexity at each layer, i.e. divide-and-conquer.
  2. Protocols can be designed and built independantly of other layers, which simplifies maintenance, allows competitive solutions and can simplify the standardisation process.
A well-recognised drawback of layered architectures is the performance compromise made to ensure clear separation of layers. This is of particular concern when resources are limited, such as in wireless and ad hoc networks. This has led to the study of cross-layer design and cross-layer optimisation principles in such networks.

A well-known communications architecture is the Reference Model for Open Systems Interconnection (OSI), standardised by ISO and ITU-T. The seven layers, shown in Figure 1, each perform individual functions, but together provide a communications service to the end users.

Layered Communications Architectures
The OSI and Internet layered communication architectures

The Internet

The prominence of the Internet in the past decade has given rise to a basic communications architecture centred on the Internet Protocol, IP (see Figure 1). IP is the core protocol in the Internet; it provides a common protocol that all applications can use, and can run over a many different network technologies.

On top of IP are the standardised Internet transport protocols: TCP, UDP and SCTP. Of particular importance is TCP, the Transmission Control Protocol, which provides a reliable end-to-end connection between two computers. The performance of TCP is highly influential on the performance perceived by users of many common Internet applications.

Application protocols such as HTTP, FTP and SMTP allow interchange of information using one of the many networked applications (Web browsers, file download, email).

In all layers of a communications architecture, including the Internet architecture, consideration must be given to several key factors which influence the success of the communications service:

  • Performance: Specific hardware technologies provide raw performance (bit rate, delay etc.); network protocols and architectures must be designed to introduce as little as possible overhead to this performance.
  • Priority: In many cases, some users require priority over other users of the network. Quality of service control mechanisms are need to ensure the correct priority is applied to network traffic.
  • Security: Protocols and architectures must provide an adequate level of security for the intended use. Note that although network technologies may be the same, in many cases the security requirements differ (e.g. a military wireless network must be more secure than a community wireless network).
  • Maintainability: Design trade-offs result in the often need for upgrading and replacing protocols. Networks must be designed with this in mind, such that the transition to future, improved systems is easy.
  • Reliability: Communication networks are depended upon in many aspects of life (e.g. financial transactions, emergency operations, business communications). The networks must provide a satisfactory level of reliability for the intended use.

Key Research Challenges

Network architectures and protocols must be designed to suite the diverse demands of users' applications, as well as the wide range of network technologies available today, and in the future. Key challenges that must be addressed in the future include:

  • Design and optimisation of protocols for new access technologies and new applications
  • Ensuring current protocols can be seamlessly upgraded to new protocols (e.g. IPv4 to IPv6 transition)
  • Algorithms for ensuring Quality of Service control can be provided across heterogenous networks
  • Securing the end-to-end communication path, while maintaing satisfactory performance
  • Analysing the impacts of cross-layer design and optimisation in existing and future networks
  • Designing network architectures and protocols for specific scenarios, such as mobile networking, ad hoc networks and satellite networking.

More Information

Australian Networks and Protocols Researchers

Researcher
Abolhasan, Mehran
Ardon, Sebastien
Bean, Nigel Geoffrey
Blackmore, Kim Louise
Boulis, Athanassios
Chan, Terence Ho Leung
Chiera, Belinda Ann
Conder, Phillip
Dadej, Arkadiusz (Arek) J
Esselle, Karu P
Fu, Qiang
Gao, Jason
Gondal, Iqbal
Herborn, Stephen Robert
Hinton, Kerry
Huang, Qing
Jamalipour, Abbas
Jayasuriya, Aruna U
Jones, Haley M
Kellett, Christopher M
Khan, Jamil Yusuf
Kibria, M. Rubaiyat
Landfeldt, Bjorn Gustaf
Lee, Ivan
Libman, Lavy
Lowery, Arthur James
Maennel, Olaf Manuel
Moors, Tim
Nadarajah, Nishaanthan
Perreau, Sylvie L
Portmann, Marius
Rakotoarivelo, Thierry
Ros, Montserrat Beverley
Roughan, Matthew
Rumsewicz, Michael Peter
Sakhaee, Ehssan
Sekercioglu, Ahmet
Seneviratne, Aruna
Shen, Gangxiang
Shu, Feng
Sun, Jinsheng
Thanabalasingham, Thayaparan
Tian, H
Tian, Shuang
Yi, Xun
Zaidi, Zainab Razia

Note: You can search for ACoRN Members using the Member Search facility