Routing In Delay Tolerant Networks Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Wireless networks refers to the network in which communication is take place in air in the form of radio signal, infrared, Bluetooth instead of cable as in wired network. A wireless sensor network (WSN) consists of spatially distributed autonomous sensors to monitor physical or environmental conditions, such as temperature, sound, pressure, etc. and to cooperatively pass their data through the network to a main location. [1]

DTN (delay- tolerant networks) is basically deals with the occasionally-connected networks that may suffer from frequent partitioning or disconnection. The networks which are suffer with disruption and disconnection and those with high-delay comes under DTN. Sensor-based networks, terrestrial wireless networks, satellite networks with moderate delays and periodic connectivity, and underwater acoustic networks with moderate delays and frequent interruptions due to environmental factors are the types of network are the example of DTN. Architecture of DTN defines a new layer which is above the transport layer and below of application layer named "bundle layer", used for store and forward bundles or messages. This architecture can be applied to many different types of challenged environments. Challenged environment in networking refers to those networks which have constraint like- lack of infrastructure, disconnection, disruption, lack of resources.

Architecture of Delay-tolerant networks (DTN)

The architecture embraces the concepts of occasionally-connected networks that may suffer from frequent partitions and that may be comprised of more than one divergent set of protocols or protocol families [2]. The basis for this architecture lies with that of the Interplanetary Internet, which focused primarily on the issue of deep space communication in high-delay environments. DTN is an approach to computer network architecture that seeks to address the technical issues in heterogeneous networks that may lack continuous network connectivity [1].The architecture of DTN is providing a new communication paradigm in which communication is possible even if end-to-end connectivity is rarely achievable. The application of DTN mostly focused on in a scenario where end-to-end connectivity is not feasible such as the connectivity problems in rural and developing regions, interplanetary and military communication, and other adversary environments. The greatest challenges in DTN is that how should it enabled end to end communication in heterogeneous environment which containing severe performance impairments. Such impairments are delay, link connectivity, bandwidth, frequent disconnection etc [3].

DTN have some properties, which makes it to use different approach as in Internet architecture-

Intermittent connectivity

In the DTN the whole network does not necessary to be fully connected at all time. Sometime some part of network temporarily isolate with the network and connect whenever find opportunity (hence called opportunistic network). This type of connectivity called as intermittent connectivity.

Store and Carry approach

For overcoming the problem with the DTN it uses Store and Carry approach in which the DTN router needs persistent storage for their queues. Disconnection, delay, flow control and retransmission etc are the reasons for applying this approach in DTN. In this approach node or custodian keep the data till the right communication opportunity will be available and then passes data to next node and so on.

Bundle layer

DTN architecture introduces a new bundle layer [4] added between the application layer and the transport layer. This bundle layer implements a store-and-forward mechanism, in which a node can store and carry data in its own buffer, and forward these data to other connected nodes whenever they are available. The basic unit of data in this layer is called bundles or messages. DTN bundle layers communicate between themselves using simple sessions.

Variable-length Messages

DTN architecture uses variable length messages for enhancing the ability of network by make good route selection decision when possible.

DTN architecture may be used in different application Area:

Interplanetary Internet

It is a computer network in space, consisting of a set of network nodes which can communicate with each other. The Interplanetary Internet is a store-and-forward network of internets that is often disconnected, has a wireless backbone fraught with error-prone links and delays ranging from tens of minutes to even hours, even when there is a connection [5].

WSN with DTN architecture

Wireless Sensor-based Networks in which nodes operate at low power, often with weak or intermittent radio communication is follow DTN architecture. In these types of networks it has some node which act as DTN nodes, node in which bundle layer implemented.

Terrestrial Mobile Networks

These are the networks which connect mobile devices and cannot ordinarily maintain end-to-end connectivity. These networks may become unexpectedly disconnected due to node mobility or variable signal strength, while others may be partitioned in a periodic, predictable manner like satellite.

Military Ad-Hoc Networks

These networks work in hostile environments where factors like- mobility environmental, intentional jamming may be cause for disconnection. As an example in reconnaissance operation using (Unmanned Aerial Vehicle) where the goal is to monitor a specific area and to report all such activities happens at that particular area [6].

Classification of Routing Protocol

Routing Protocol is the rules to selecting and searching the paths in a network from which to send and receive network traffic. Routing protocols in the wireless network is different than in wired network and affected by various factor. Routing can be divided into these two classes as follows-

1-Pro-active routing

It also called table driven routing. In this type of routing the information about route and position of nodes are predefined and kept in the form of table. Hence the optimal route can be found with minimal delay but require enough resources to keep these tables up to date.

2-Re-active routing

It is also called on demand routing. In this approach it does not need predefined information about route. This type of technique finds a route on demand by flooding the network with Route Request packets but incur high latency time in route finding.

Routing Protocol in DTN

Randomized routing

In these scheme packets randomly jumps around the nodes until it finds the destination. The nodes sending the packets to its neighboring nodes, so after some time the packet reach to the destination node and stop sending the packet afterwards. At best case it will reach to the destination in one jump or one hop distance and at worst case it will have to access all the nodes present in the network.

Epidemic routing

In these scheme the all the nodes of the network receive the copy of the packets. The sender node flooding the networks by the packets until the all nodes has the copy of the packet. This routing approach just likes to transfer the data in the distributed database where one copy of data present in the entire database.

It can also classify routing protocols in DTN based on the number of copies of the same message that are created


In this scheme, a single-copy message is forwarded through multiple intermediate nodes to the destination.

Quota replication

A quota-replication scheme makes a specific number of copies for a message as specify by a quota [7]. Quota decides the maximum number of duplicate messages of any packet store in the buffer of node.


A flooding scheme makes an extremely large number of copies and spreads them throughout the whole network for communication.

Geographical Routing

The routing in the wireless routing taking into accounts of the topology of the network and changes as if the topology changes. So the network in which topology changes frequently need more overhead in routing schemes. So what if it pays attention only on the physical or relative location of the nodes, then the routing is just based upon the position of the nodes. So the routing principle which relies on geographic position information called geographical routing [8]. The most famous approach under geographical routing technique is greedy approach [9] but it is not always successful so it needs some backup approach also.

Location Service

For geographical routing, it has to need a service or mechanism which provides it the current geographical location of the destination for the packets and is called location service. In geographical routing it often chose the service take advantage of proactive element of routing. The location service of that kind can be divided into the two parts

Dream location service

A node shares its position to nearby nodes at a defined given rate and to nodes further away with lower rate. The rates depend on a node's speed.

Simple location service

A node exchanges its position to only to its neighbors. So that exchanging location tables between neighbors, Communication is kept local while permitting the location data to globally be distributed in the system.

Issues in DTN

Limited resources

Networks using DTN architecture often deploy in an environments which are not generally suitable for networks using Internet architecture. So that in such environment the human intervention is very less. It has to pre-decide what and how it provides resources to network for its optimal working. So the balance among memory, power, bandwidth and processing capabilities is necessary.

Low data rate, high latency

The delay is the major characteristics of the DTN, so that some time due to disconnection or disruption packet would store in DTN nodes (bundle layer) so its incur overall network high latency and low data rate.[10]


In DTN frequent disconnection is major issue. The overall design solution of such networks takes mainly disconnection as major problem.


Networks use DTN architecture in which end-to-end connectivity is merely available and sometime disconnection last for a long period of time. So that it needs enough amount of storage to store these bundles.


In many circumstances that some form of authentication and access control to the network required. An unauthorized user can flood the networks so denying service to authorized users and unauthorized user can access the mission-critical private data etc. So it needs to pay attention on the security and integrity of data as well as to the networks.

Environmental disruption

DTN architecture is meant for the networks having different environmental condition like- Interplanetary networks, underwater acoustic network, and military Ad-hoc network etc. All of these are have different environment and different issue like- no medium available, communication underwater, noise etc.

Related work

The one of the area of research work in DTN is efficient routing on challenged network. Authors [6] proposes LAROD-LoDiS scheme Applicable in Intermittently Connected MANET and implement it in reconnaissance mobility using UAV (unmanned aerial vehicles). The LAROD-LoDiS scheme is more effective and efficient compared with a leading non-geographic scheme spray and wait. It is kind of geographic routing with store-and-forward principle. It first uses greedy approach to forward and when this approach is no longer working it holds or store the packet until node mobility become available. DTN sender node first broadcast the message, all the DTN nodes within the pre-specified area is eligible to further forward the message. All DTN nodes set a delay-timer t and whose timer expires first choose as the new next DTN node. This delay timer set and chosen by Delay-Timer function and like this it finds destination node. When the destination receives the message, it generates an acknowledgement (ack) and floods it in network. All nodes which receive acknowledgement store it until the TTL (Time -to-Live) of packet expires. In the reconnaissance scenario, spray and wait fails to provide an acceptable delivery ratio within a reasonable delay, whereas LAROD-LoDiS can provide more than a 95% delivery ratio [6].

Authors [11] proposes another geographic routing scheme for overcoming the challenge to obtain the real-time location of destination in DTNs. In this paper they first calculate the proximity to the movement range estimated for destination by considering the mobility vector of mobile node. Then the message is replicated by the algorithms named as Reach Phase and Meet Phase respectively, for the sake of fast message delivery. If the message is out of the movement range then it use the Reach phase and if it is within the range then it use Meet phase. Apart from this, it calculate anti-diffusion function to reduce the redundancy. The replicated messages are again under prioritized transmission with the help of scheduling scheme between Reach and Meet Phase, for enhancing the routing performance [11]. The proposed algorithms achieve the highest delivery ratio comparing with POR (Packet Oriented Routing) and DAER (Distance Aware Epidemic Routing).

Authors [7] proposes an algorithm to reduce the network congestion. This mechanism can be applied to any routing algorithms which are based upon quota replication. Where quota decides the maximum number of duplicate messages that packet stored in the buffer of nodes. In this paper, they combine a quota-replication routing technique and a message moving technique to provide solution for congestion control. Performance evaluation shows that this proposed solution outperforms other congestion controls mechanism [7]. In this paper they describe the mechanism to control the number of copies introduced into the network based on congestion condition. So that they it makes a balance between high delivery rate and low drop rate. In this paper every node contain two tables- NBOT (Neighbor Buffer Occupancy Table) and CPT (Contact Probability Table).NBOT records the buffer occupancy of each neighbor nodes within its range and CPT records the contact probability of each node which has been encountered before. And based on this they choose their routing decision.

Authors [12] proposes a modified and improved version of "PRoPHET " algorithms which is a routing protocol for DTNs that was developed when DTN research was in its conception and which has been studied by many afterward. PRoPHET uses the history of previous encounters with other nodes as well as the transitive properties of the network to optimize the bundle forwarding over the network by pruning the epidemic distribution tree [12].it calculate delivery predictability (DP), which is use for predicting the node encounter timestamp for network. More the value of delivery predictability (DP), more recent information it may have. Because of the intermittent properties of the network some node only connect for few time but have greater delivery predictability (DP) value .hence they modified the old PRoPHET algorithms by few changes. PRoPHET-v2 performs better than the old PRoPHET routing protocol, in cases of heterogeneous network mobility scenarios especially in the simulation results.

Authors [13] proposes a solution for routing to the WSN (Wireless Sensor Networks) which is similar to the DTN routing approach using store-and-forward mechanism. The node which stores the data called "Mobile Sink". There can be one sink or multiple sink to record and manage the data behalf of the nodes presents in the network. They do not differentiate the individual mobile sinks from one another; sensor nodes can forward data to any of them, since all the sinks forward the data further on to the gateway of the network, which then process it. The sink in the network are responsible to collect and manage data from sensor nodes, so the node which are neighbor of the sink used so frequently as all routes passes from this nodes. So that these node's battery power and resources used frequently than others so they are the first nodes which are affected. So for overcoming from this and for increasing lifetime of the network we use mobile sink, so that we can optimize the network. Mobile sinks may move on a fixed path may take a random walk or may move as instructed to optimal walk in terms of network lifetime. The basic principle of this routing protocol is to calculate a cost between the sink and each sensor node, and store it in a table. The cost function may choose arbitrarily based on factors like- energy of node, distance of node etc. If node is not reachable by node then it starts discovery service to find new rote or until new route become available. [13]

Authors [14] proposes the flooding-based routing schemes for using in intermittently connected mobile networks (ICMN) which is one of the types of Delay Tolerant Networks. Although flooding based mechanism have high probability of delivery but it also have some issues like- they waste a lot of energy and suffer from severe contention, which can degrade their performance. This routing scheme "Spray and Wait" it sends a number of copies into the networks known as "Spray" and then waits till one of these copies reaches the destination known as "Wait". This scheme is manage to overcome the shortcoming of epidemic routing and other flooding based schemes. By simulations it shown that Spray and Wait outperforms other existing schemes with respect to number of transmissions per message delivered and average message delivery delay [14].

Authors [15] proposes the current DTN evaluation practices through a thorough and literature study. Authors shows some weakness used in the evaluations and proposes a model for evaluation of DTN routing schemes that track the crucial inputs which need to be considerate in the evaluation process. Authors first summarize the related work then chronological overview and then proposed DTN evaluation model. The model took three classes of inputs to generate the results in the simulation environments. In these three classes the first one is take care about connectivity, mobility, network traffic, node resources, second one consider routing protocol, protocol parameters and third class is about time scale. These three classes evaluated under simulation environment and produces results which include delivery rate, hop count, delivery rate etc. These evaluations infer that DTN routing requires to consideration of all type of input into simulator to understand the routing in all aspects. [15]

Authors [16] proposes gradient-based routing method in the Delay Tolerant Mobile Sensor Networks (DTMSNs) that incorporates node mobility into its gradient to reduce the effect of inefficient message relays. According to author this scheme reduces the average message delivery delay by up to half in the case that nodes move straightly. There are two types of routing methodology in the DTNSNs as- single copy, multi-copy. Multi-copy forwards multiple copy of packet in the networks to hope of one of the copies successfully arriving at a destination node. In single-copy routing, each node calculates a value that defines how useful a node might be in forwarding messages to the destination node and transmits messages based on this value. When node communicates with other node it forwards the message to the node with the highest value. This kind of approach is also called Gradient-based routing. This paper describes how node mobility affects the message delivery in gradient-based routing and proposes a gradient-based routing method that incorporates node mobility. By simulations, it showed that algorithm improves the average message delivery delay compared to the existing gradient-based routing method. [16]

Authors [17] proposes DTN routing algorithm for urban public transport systems. The need for the urban public transport systems is for Vehicle dispatching, traffic management, arrival prediction, and dynamic passenger information. The algorithm trying to exploits the characteristics of public transport systems to achieve a better efficiency. Public transport systems have some specific characteristics as- timetable and network map which must be exploited in order to achieve better routing. The algorithm work well in higher bundle load where other algorithms did not performed well but simulation result shows that this mechanism routing in Urban Public Transport Systems (RUTS) work efficient only when timetables are known previously.[17]

Author [18] proposes a framework for examining the different routing performance in DTN. This framework uses for deriving the exact packet delivery delay and communication cost distribution for a large class of delay tolerant network routing protocols. The authors give the first theoretical framework for examining packet delivery delay and communication cost. It is also examining the epidemic routing performance at different node cooperation within the network. The proposed framework is based upon the characterization of the routing process as a stochastic coloring process. They takes assumption in their proposed framework which are- low load, no contention, fast transmission and two node communicate with each other only if they are within a transmission range and encounter happen only when they reach into their transmission range. Framework implements their application as example to routing protocols namely epidemic and two-hops routing.[18]

Comparison among existing protocol

These proposed algorithms perform better to specific condition and have some limitations. The algorithm in [6] is implemented in reconnaissance operation using UAV (unmanned aerial vehicle) for military operation. In these algorithms the delivery ratio is improved as the node density is improved but it does not work well for sparse node systems. It also has to maintain the clock synchronization and it may be possible that the geographic position of the node is corrupted or may be very old. The algorithm in [11] the real-time location is estimated by GPS (global positioning system) without an error which is not true. It uses geographical routing scheme which need real-time location of the node which is hard to get in high mobility network. The algorithm in [7] deals with the congestion and algorithm in [12] is the improved version of the PRoPHET algorithms. But somehow these two algorithms work same as both of these algorithms exchange data from other node and modify entry on its own table. And based on these data they decide and make their routing decision. So these two algorithms take their basic idea from the PRoPHET algorithm. Algorithm [13] proposes a solution for routing to the WSN which is similar to the DTN routing approach using store-and-forward mechanism. The node which stores the data called "Mobile Sink". The mobile sink is similar to the DTN link where persistent storage will need. Algorithm [14] is one of the main routing algorithms of DTN routing called-Spray and Wait. This is using flooding based approach for high probability of delivery but it suffers from wastage a lot of energy and severe contention, which can degrade their performance. Authors [15] shows some issues used in the evaluations and propose a model for evaluation of DTN routing schemes that track the crucial inputs that need to be considerate in the evaluation process. The model is proposed for simulation environment. Algorithm [16] is a gradient-based routing method in the Delay Tolerant Mobile Sensor Networks that incorporates node mobility into its gradient to reduce the effect of inefficient message relays. Like PRoPHET in this scheme node also calculate the value and forward the packet to the node which has highest value. Algorithm [17] is a DTN routing algorithm for urban public transport systems. For achieving efficient routing in that type of networks it have to exploit these two - time table and network map. This algorithms work well but when timetable is pre defined. And at last a framework [18] which examines the different routing performance in DTN. It is the first theoretical framework for examining packet delivery delay and communication cost. Although it is a good framework but it has some assumption about network like- low load, no contention, fast transmission and two nodes communicate with each other only if they are within a transmission range.

Proposed problem

All these algorithms use geographical routing technique to determine the location of destination and the next forwarding node. To determine geographic location they rely heavily on the devices like- GPS or some techniques like beacon based to find relative distance. Both these techniques have some inherent error for exact location so need some improvement techniques for location scheme. To efficient working of the node we have used it as optimal as possible. For doing it so we need a better tradeoff between the space and time for achieving better performance. For the networks of long delay we have to carefully decides TTL for the bundles.


The DTN architecture is use in the heterogeneous networks that must operate in environments subject to long delays and discontinuous end-to-end connectivity. Now there are various areas which are using DTN architecture like- Interplanetary Internet, sensor networks, MANET and VANET, terrestrial wireless networks, satellite networks with moderate delays and periodic connectivity, and underwater acoustic networks etc. This architecture has some challenges like- storage, delay, disconnection, limited resources, security etc. The on-going research is based on to further optimize this architecture and work upon its challenges.