Research Article
Spray and Wait Routing With Agents in Intermittently Connected MANETs
Regional Centre of Anna University, Madurai, Tamil Nadu, India
P. Ganesh Kumar
KLN College of Engineering, Sivagangai, Tamil Nadu, India
Mobile Ad Hoc networks (MANET) (Shi et al., 2010) is the hasty exploitation of autonomous mobile users. Substantial example (Kuiper and Nadim-Tehrani, 2011) includes establishing survivable, efficient, dynamic communication for emergency/rescue operations, disaster relief efforts and military networks. Such network setups cannot count on centralized and organized connectivity and can be conceived as the applications of MANET.
A MANET is an independent anthology of mobile users that commune over relatively Band Width constrained wireless links. Due to the node mobility, the network topology may change swiftly and erratically over time. In quintessence, the network is decentralized; where all network activity including discovering the topology and delivering messages must be executed by nodes themselves, i.e., routing functionality will be assimilated into mobile nodes (Kuiper and Nadim-Tehrani, 2011).
The routing methodologies in such networks are made through traditional routing protocols like AODV (Ad hoc On demand Distance Vector) (Lakshmi and Sankaranarayanan, 2006, Asfand-e-Yar and Sher, 2004), DSR (Destination Source Routing), Link State routing protocol (Alagiri et al., 2012) On demand Multipath Routing (Ramesh and Kumar, 2012a)etc. Ample routing (Wang et al., 2008; Bhagyaveni and Shanmugavel, 2005; Manickam and Shanmugavel, 2007) algorithms have been proposed for MANET since the past two decade. The mobile Ad Hoc network (Sesay et al., 2004; Nazir et al., 2006) drawn from wireless sensor networks (Ramesh et al., 2012a) paves to a new form of network called the Intermittently Connected Mobile Ad Hoc Networks (ICMANET).
ICMANET is meticulous to be one of the new areas in the field of wireless communication. This form of networks araise due to the high mobility of nodes in the network that tremendously leads to distorted network. ICMANET, also known as the Delay Tolerant Network (DTN) (Stamouli, 2003), is typically different from traditional MANETs (Sharma et al., 2007; Ramesh and Kumar, 2012b) which means that in the latter; communication between two nodes is possible at any time via a path of intermediate nodes although this path may vary with time. The main idea behind ICMNAET is that a complete path between nodes could never exist at any point of time during the process of network communication.
The traditional routing scheme that forms a basis for the routing schemes in ICMANET is the Flooding based routing. In this, one node sends packet to all other nodes in the network. Each node acts as both a transmitter and a receiver. Each node tries to forward every message to every one of its neighbours (Cokuslu and Erciyes, 2008). The result in every message eventually is delivered to all reachable parts of the network (Khan et al., 2008).
Following this, the Epidemic Routing Protocol (Vahdat and Becker, 2000) was ensured for routing in ICMANET but it does not exert such an effect due to its pair-wise connectivity. It is simple in nature. Spray and Wait outperforms Epidemic in characteristics like minimum delivery ratio and flexibility. This study speaks about the means of secure communication in ICMANET. In the following sections, the routing methodology and agent setup with various parameters are discussed.
SPRAY AND WAIT ROUTING PROTOCOL
The Spray and Wait routing protocol (Spyropoulos et al., 2005) oeuvres in core of spraying a few message copies into network and then waiting for them to get in contact with target node in concord to its name. In general, the Spray and Wait spreads a scarce amount of message packets into the network and waits for a bounded amount of time until one of the nodes in the network get in contact with the destination node.
The Spray and Wait (Kuiper and Nadim-Tehrani, 2011) outstrips the other traditional routing schemes with its average message delivery ratio, number of transmission hops to reach the destination on packet delivery and the overall routing performance. The implementation complexity of Spray and Wait (Spyropoulos et al., 2008a) is undemanding and it also can be heightened to pull off the performance in certain. It is palpable that this routing scheme limits the amassed number of transmissions per message packet without conciliation with the performance.
The Spray and Wait protocol holds two phases namely the Spray phase and Wait phase.
Spray phase: For every message originating at a source node, L message copies are initially spread forwarded by the source and possibly other nodes receiving a copy-to L distinct relays.
Figure 1 depicts the L message copies initially spread to L distinct nodes. The spray scheme is done in two forms either the source spray or binary spray and wait scheme. Both schemes are explained subsequently (Diana and Lochin, 2012).
Fig. 1: | L message copies sprayed |
Fig. 2: | Source spray mechanism, all copies of message transferred to all L distinct relays |
Fig. 3: | Binary spray mechanism |
Source spray: Source spray is the simplest approach in which the source node forwards all L copies to the first L distinct nodes it encounters.
It means that the source node hand over all the packets that were generated in it to the node that it has encountered within its radio range. This is clear from Fig. 2.
Binary spray and wait: The source of a message initially starts with L copies; any node A that has n>1 message copies (source or relay) and encounters another node B (with no copies), hands over B n/2 and keeps n/2 for itself; when it is left with only one copy, it switches to direct transmission. The process is portrayed in Fig. 3.
Wait phase: If the destination is not found in the spraying phase, each of the L nodes carrying a message copy performs direct transmission (i.e., will forward the message only to its destination).
Once during spraying phase (Spyropoulos et al., 2008b) if the destination node is not found, each of the L node that holds the copy of message packet performs the direct transmission as it encounters the destination and to any other relay nodes in the network.
AGENTS IN SPRAY AND WAIT ROUTING
The Agent is a program module that functions incessantly in a meticulous environ. It is proficient in carrying out activities in a supple and intellectual comportment, that is responsive to changes in the surrounding environ. The agent is not a complete program but is an interface (Beson and Leleu, 2009) responsible for performing the pre assigned chores. Agent is autonomous which takes actions grounded on its innate knowledge and its precedent experiences (Ioannis et al., 2007; Ramesh et al., 2012b).
On setting agent (Ramesh and Kumar, 2012b; Ramesh et al., 2012c) at each node, security is achieved by incorporating certain agent parameters at each node. The agent parameters are described as follows:
• | Node ID: A unique identifier for each node in the network |
• | Passcode: A common password for the nodes in the network |
• | Mobility model: The mobility model of the network topology |
• | Origin of placement: The initial placement of each node in the network |
• | Grid card: An nxn matrix in which each grid contains a particular data in it. Each node contains a unique grid |
• | Pattern formation: The network is associated with certain geometric silhouettes and using these, each node encompasses a unique pattern |
These agent parameters settle on the security issue coupled in the ICMANET.
Agent is set in each node and it includes three components (Hegazy et al., 2003) namely as follows:
• | Data aggregator: The data aggregator is similar to a database that holds an aggregate of information about all the nodes within the network. It includes detailed information of each and every node. The aggregator holds the agent parameters of every node. In plain, it is just the collector of information |
• | Node analyzer: The node analyzer analyses whether the node that accept the information is a family node i.e., node that belong to the topology. The analysis is based on the agent parameters that are hoarded in the data aggregator. It selects any one of the parameter in a random manner to conclude that the node is a family node. It works as a typical form of intrusion (Kachirski et al., 2002; Stamouli, 2003) detection mechanism. If a malicious node is sensed, the node analyzer broadcasts the presence of intruder |
• | Data broadcaster: The data broadcaster, once after a node is determined to be a family node, it allows the sender or any relay node to transfer the message packet to the secondary relay node. It acts as a gateway that provides access for communication amidst the encountered nodes |
From Fig. 4 the working of agent set at each node. When a secondary node receives a message packet from a primary node, the node analyzer analysis whether the node belong to the restricted network. The node analyzer selects one of the parameter randomly and checks for the authorized node from the data aggregator. The data aggregator is a database, if a match is found within the data aggregator; it is preceded towards the data broadcaster. If the node is not valid, node analyzer broadcasts the presence of the intruder within the network. The data broadcaster allows the node to transfer the message packet to the encountered node.
Fig. 4: | Working of agent |
SIMULATION RESULTS
This section describes the simulation results of spray and wait with the setup of agent in it. The spray and wait routing protocol with agent has been evaluated using one simulator (Keranen et al., 2009, 2010). The normal routing and routing with agent have been evaluated and compared to attest the performance criterion of the protocol. The following sections clearly show the scenario setup for the Spray and Wait (SNW), the performance of Spray and Wait and the performance of Spray and Wait with Agent (SNWA) are elucidated. Subsequent to these the performance of routing with and without agent are compared.
Scenario setup: The parameters set are the basic one simulator (Keranen et al., 2009, 2010), environ parameters and are given in Table 1. The one simulation for Spray and Wait, in this study, uses the random waypoint model. The nodes move in an area of 2000x2000 m with a speed limit within bounds 0.5-1.5 m sec-1. The radio range is set to 250 m. The efficiency of any routing protocol is determined by the node density i.e., the total number of nodes within the set network.
The packets are generally generated with the initial setup of the simulation and holds through the overall simulation time. The Time to Live (TTL) or the packet life time is set as 600 sec initially that are varied lately on consideration to the performance criterion. When evaluating, the simulation is run for 3000 sec.
Performance of SNW: The performance of SNW is greatly influenced by the number of nodes, overhead and the delivery delay (Subramanyam and Prasad, 2006).
Table 1: | Basic simulation parameters |
To evaluate the performance of SNW routing, the number of nodes are varied with respect to various network parameters like overhead, delivery probability and latency. The test continues by varying packet lifetime followed by distinct transmissions per message.
Performance of SNWA: On setting of agent the performance of Spray and Wait to route packets in ICMANET should not degrade. As setting of agents eventually decreases the time complexity i.e., it decreases the time required to route packet. The agent setup ensures a higher scale of security. These general properties of Agent along with SNW Routing, endow with a secure communication or transfer of message packets across the network. An optimal result with agent setup at each node is simulated.
The performance of any routing protocol is assessed by the node density, node speed, its life time in disparity to which the overhead, number of transmissions, delivery ratio and delay are to be maintained optimal. The setting of agent provides high security. This study shows a better performance amidst these concerns.
To evaluate performance, the various network parameters were evaluated by varying the number of nodes, packet life time or packet generation time. It is shown that with agent, Spray and Wait shows optimal result.
Performance comparison of SNW and SNW agent: To show the pro of spray SNWA, the simulated result of both SNW and SNWA are compared. SNW is an efficient routing algorithm with good delivery probabilities. In order to mark the outperformance of SNWA, a comparative analysis is made between SNW and SNWA grounded on some basic network parameters.
The comparison is made using various metrics as follows:
• | Number of nodes vs. delivery ratio |
• | Number of nodes vs. overhead |
• | Number of nodes vs. latency |
• | Transmission range vs. overhead |
Number of nodes versus delivery ratio: Comparing the delivery ratio i.e., the probability to deliver the message, initially SNW provides an optimal delivery ratio. On setting of agent, it provides considerably slight higher ratio. Both SNW and SNWA vary with slight variation in delivery ratio and are shown in Fig. 5 with respect to the number of nodes.
The Fig. 5 shows that on setting of agent, SNW shows slight modification in its performance. The SNWA exerts a minimum of 33% (aprox.) higher delivery probability than SNW protocol. Both schemes deliver with the average probability that is required for an efficient routing whereas SNWA provides higher delivery ratio and secure routing.
Fig. 5: | Delivery ratio SNW and SNWA with respect to numbers of nodes |
Fig. 6: | Overhead for SNW and SNWA assorted number of nodes |
Number of nodes versus overhead: The overhead remains a major network parameter need to be considered to evaluate the performance. The agent completes the task within a few milliseconds than SNW in accordance to its basic characteristics. SNWA exerts an overhead which is 6% (approx.) lesser than SNW.
Hence, the performance during higher node density is considered to estimate the performance of SNW on setting of agent. The comparative result of overhead in SNW and SNWA are depicted in Fig. 6 with respect to the number of nodes. From the Fig. 6 it is apparent that SNWA exerts a minimum overhead as it reduces the time required to communicate. The result proves that SNWA is better in contrast to SNW.
Number of nodes versus latency: The time criteria once again stand forth as an important aspect during the measure of latency. The setting of agent takes a few milliseconds less to route. This is because of the basic nature of agent to use lesser time. Figure 7 shows the comparative analysis of latency in both SNW and SNWA and the analysis is made over average latency with respect to the number of nodes set in the network.
The delay in SNWA is 2% lesser than SNW and shows a better performance. This variation in delay is mainly due to the time reduction characteristic of agents. With lesser delay SNWA renders secure message transfer. The secure and efficient transmission is evident from the Fig. 7.
Transmission range versus overhead: The transmission range is referred to the radio range within which each node can communicate with other node. SNWA shows a better performance even when the transmission range is varied. In this, the overhead ratio is evaluated in contrast to the transmission range. The evaluation of SNWA and SNW with respect to transmission range and overhead is portrayed in Fig. 8. As the transmission range increases, the probability of node interaction decreases, even in such scenario SNWA outperforms SNW, satisfying the basic properties of SNW.
Fig. 7: | Delivery latency SNW and SNWA for distinct number of nodes |
Fig. 8: | Overhead for SNW and SNWA with respect to transmission range |
From Figure 8 it shows that the overhead in contrast to the transmission range, SNWA illustrates a variation in the range of 11% (approx.) lower than SNW. Overhead decreases the efficiency of routing protocol and is considered to be the important parameter. In this paper, SNWA depicts a better overhead performance with its quality of doing task in fraction of second and providing higher degree of security in performing the tasks. Spray and wait generally exerts optimal delay with minimum overhead. On setting of agents the routing performance varies approximately in the range of 40% with higher degree of security.
In this study, the optimized performance of SNW Routing protocol with agent is demonstrated. On setting agent this protocol ensures a secure delivery of the message packets. It is evident from this that the basic properties of SNW protocol like fewer transmissions, low contention, better delivery delay and high scalability are achieved with a generalised value. A comparative analytic result of SNW and SNWA roofed in this paper. From this it is obvious that SNWA outperforms SNW with an approximate ratio of 33, 6, 2 and 11% with respect to delivery ratio, overhead, delivery latency and transmission range. This paper, in general ensures secure communication without any performance degradation. This paper proves secure routing with setting of agents.