Delay Tolerant Networking

The Delay - Tolerant Networking (English for delay- tolerant network, also called Disruption - Tolerant Networking, for non-disruptive tolerant network, and shortly DTN ) is a protocol architecture to overcome the technical difficulties associated sparse and heterogeneous communication networks. The architecture is based on the model developed by the U.S. space agency NASA interplanetary Internet (IPN ), generalizes this design however. In contrast to the design focus of the IPN - the space communications - focused on the DTN networks with low stability of end-to -end connection.

Development

Around the same time responsible for the development of mobile ad hoc routing method that DARPA in the 1990s regardless of NASA, MITRE and others to develop a proposal for an Interplanetary Internet (IPN ). A basic design for an IPN architecture was developed by a group led by Vint Cerf, taking into account the susceptibility to interference of network connections (Packet Corruption ) in space and the necessary treatment long message delay times (packet delay) presented.

In 2002, Kevin took the case on some of these ideas for the development of a general disruption tolerant network and put his design in 2003 under the title Delay Tolerant Networking and the derived therefrom Acronym DTN on the SIGCOMM conference before. In subsequent years, the conditions for an appropriate network at various conferences were discussed, which can also deal with long signal transit times and disconnects. Here, the experience with sensor networks and mobile ad - hoc routing methods should be considered. The routing algorithms and procedures for ensuring the reliability and verifiability of the data transmitted could be improved over time.

Routing

The transmission and routing of information to be provided information from the data source to the target are a basic necessity in all data networks. Here, DTNS differ primarily by the lack of permanent connection between data source and data target of other data networks. Ad- hoc routing methods, such as AODV and DSR, could not be used under these conditions, because they are trying only to determine the full path from the source to the destination, and only then transmit the data. However, if a continuous end-to -end communication can be difficult or impossible to ensure the approach of the store and forward principle is meaningful. The data in small packets from one network subscriber is transferred to the next available and each is stored until the transfer to the target computer was successfully confirmed. Transmitting the message in the same version of a plurality of paths from the source to the target is a common approach to increase the probability of successful message transmission.

The RFC 4838 and RFC 5050 were published in 2007, they provide a necessary for the development of algorithms and applications overview of requirements for the used in a DTN software. This is generally referred to as a bundle protocol protocol defines a series of contiguous data blocks as a bundle, each bundle containing this as opposed to an individual block of data sufficient semantic information to continue an application. Bundles are transmitted between the nodes connected to the network with the store and forward principle, as well different transport protocols can be used, which may be based on the IP protocol, but it does not presuppose. The log levels that transferred the bundles on their local network are referred to as bundle convergence layers. The bundle architecture works as an overlay network, it uses an additional name architecture based classification to Endpoint Identifiers ( EIDs ) and on a rough class of service.

The bundle protocol must balance the different needs of individual applications for sending bundles over the network. Due to the Store and forward nature of DTN protocols benefit from routing solutions that information is known by the application layer for routing. For example, consider if an application is reliant on particularly fast, obtain data in a transfer of or with constant packet delay the network.

Bundle protocols collect application data in bundles that can be transmitted over various networks with a high level of service priority. The service priority is generally dictated by the application, the RFC 5050 Bundle Protocol Specification provides for the bulk priority levels, normal and expedited ago.

Security

The safe is addressing one of the key points of the bundle protocol.

The security requirements for a DTN network differ depending on the application and working environment, but authentication and confidentiality often play a major role. To guarantee a certain security in a network without permanent connections is difficult because so cryptographic protocols or key exchange is expensive, and the need that any of the network nodes must constantly recognize other network nodes only temporarily in the network achievable secure. The implemented solutions were often derived from the mobile ad - hoc network techniques and have been influenced by research on data security, such as the use of distributed CAs and PKI schemes. From the DTN area itself applying the ID-based cryptography, which enables network nodes, with their public ID to receive encrypted data originates.

Research approaches

The differences resulting from the requirements of the DTN problems are currently being studied by various institutions

  • The Delay Tolerant Networking Research Group.
  • Technology and Infrastructure for Developing Regions project at UC Berkeley
  • The KioskNet research project at the University of Waterloo.
  • Dieselnet research project at the University of Massachusetts, Amherst.
  • ResiliNets Research Initiative at the University of Kansas and the University of Lancaster.
  • Haggle research project of the European Union
  • N4C research project FP7 of the European Union
  • WNaN DARPA project

Some research projects on DTN for an Interplanetary Internet are already investigating the use of the Bundle Protocol in Space:

  • NASA JPL 's Deep Impact Networking ( DINET ) experiment on board the space probe Deep Impact / EPOXI.
Ad hoc On-Demand Distance Vector Routing Dynamic Source Routing Lancaster University Deep Impact (spacecraft)
226049
de