GEAR: ROUTE SELECTION IN AD HOC NETWORK BASED ON SHORTEST PATH AND PATH STABILITY

 

Kumar Anand, Anindo Mukherjee, Vivek Jain

University of Cincinnati

(anandk, mukhrao, jainvk)@cs.uc.edu

 

Abstract

 

A key feature of sensor networks is that their nodes are untethered and unattended. Consequently, energy efficiency is an important design consideration for these networks. Motivated by the fact that sensor network queries may often be geographical, an energy efficient routing algorithm that propagates a query to the appropriate geographical region, without flooding is proposed [1]. Geographic and Energy Aware Routing algorithm uses energy aware neighbor selection to route a packet towards the target region. In this project the GEAR protocol is evaluated using simulation. The simulation is done using simjava, a process based discrete event simulation package. Through our simulations we find that energy aware routing gives better results in terms of network lifetime and delivery ratio.

 

1 Introduction

An ad hoc network is a cooperative engagement of collection of wireless mobile hosts forming a temporary network without the aid of any centralized administration or standard support services regularly available on the wide-area network to which the hosts may normally be connected. Ad-hoc networks are self-organizing network architectures that are rapidly deployable and adapt to the propagation conditions and to the traffic and mobility patterns of the networks nodes. The most distinguishing characteristic of ad-hoc networks is that the network does not require any fixed infrastructure [2].

 

To provide communication throughout the entire network, each node is also designed to serve as a router. However, the limitations on power consumption imposed by portable wireless radios result in a node transmission range that is typically small relative to the span of the network The result is a distributed multi-hop network with a time-varying topology.  In such an environment, it may be necessary for one mobile host to enlist the aid of others in forwarding a packet to its destination, due to the limited propagation range of each mobile host.

 

1.1 Key Features of Ad-Hoc Networks

 

The key features of an Ad-Hoc network can be summarized as follows:

 

1.2 Issues in ad hoc Networks

 

There are several issues that need to be considered while designing an ad hoc network. A few important ones are listed below:

 

One of the major issues that affects the performance of an ad hoc network is the way routing is implemented in a network. Routing protocols used in conventional wired networks cannot be used in ad-hoc networks due to the lack of ability to quickly reflect the changing topology. There are several issues that are dependent on the routing protocol used.

 

The limited battery life of every sensor node is a major design issue in the design of routing protocol. Hence we evaluate an energy efficient routing algorithm proposed by the authors to propagate a query to an appropriate geographical region, without flooding. The proposed algorithm uses energy aware routing to select the neighbor that has sufficient energy level and is closest to the destination for forwarding the query. The protocol can adapt its energy aware factor by varying a flexible parameter α that decides the weightage of the geographical closeness to the destination nodes, and the energy aware factor in deciding the best next hop neighbor for query propagation.

 

Some common assumptions of this paper:

A.     It assumes sensor nodes are static or immobile i.e.  Once the node has learned its location, its co-ordinates do not change.

B.     There is a central location database managed by a central entity which enables each of the nodes to discover its position.

 

In the real scenario, each node would learn its location by some kind of GPS system so the above assumptions can be made without the loss of generality.

 

2 Geographical and Energy Aware Routing

 

2.1 Principle of GEAR

Geographic and Energy Aware Routing (GEAR) technique uses energy aware and geographically informed neighbor selection heuristics to route a packet towards the target region. Within a region, it uses a recursive geographic forwarding technique to disseminate the packet. Although the energy balancing design of GEAR is motivated by sensor net applications, the protocol is generally applicable to ad-hoc networks.

 

We now describe the Geographical and Energy Aware Routing (GEAR) algorithm. As mentioned in the introduction, we are interested in routing queries to regions in proposed sensor-net applications. This routing algorithm works by dividing the area into rectangular regions. Each node has as associated cost to every region in the network. The process of forwarding a packet to the target node comprises of two steps.

 

  1. Forwarding the packet to the region. If the current node is not in the same region as target region, the node picks up the nest hop node among its neighbors, which has the lowest cost to the target region.

 

  1. Forwarding the packet within the region. If the current node holding the packet is in the same region as the target region, restrictive flooding inside the region is used to disseminate the packet.

 

Each query packet has a target region specified in one of the fields. Each node knows its own location and remaining energy levels, and its neighbor’s locations and energy levels by exchanging messages on demand basis. Following figure represents the simplified topology in our simulation. Area is divided into square regions and all nodes are contained in one region or other. A path or edge exists between two nodes if each has sufficient energy level so that it can forward packets and both are within the transmission range of each other.

 

GEAR behaves like an On-demand Routing Protocol [3]. In the case of on demand kind of routing protocols, the route discovery is on demand basis. Thus there is no overhead of maintaining a routing table in every node. The key features of such a protocol are as follows:

 

·        Route discovery is initiated only when a node wants to initiate a call.

·        No permanent power usage for periodically updating routing tables.

·        Create a lot of overhead when the route is being determined, since the routes are

       not necessarily up-to-date when required.

 

Node B
 

Node A- Source

Node B- Destination

 

Fig1 depicting the GEAR protocol

 

The cost of our neighbors is learnt on demand basis, i.e. only when we have a packet to forward. This was done to make sure we have the most recent cost of our neighbors; this we felt will also reduce the overhead of periodic hello messages in a highly populated network. Also, if numbers of regions are very large, and also numbers of neighbors are more, then storing a cost table for every neighbor to all the regions may be an overhead.

After a packet reaches the destined region we implement restrictive flooding as opposed to recursive geographic flooding in [1].  It helps in a simpler implementation. Also it is suitable in a low density region.

After a node reaches below a threshold value of energy, it becomes inactive and at that time it broadcasts its dead status to all its neighbors. We felt that if the dead status of a particular node is broadcast throughout the network, then the packets destined for dead nodes won’t be originated at all. However we have not implemented the network broadcast in our simulation and it contributed to some unsuccessful packets in the results. This can be used for future improvements.

 

2.2 Energy Aware Neighbor Computation

Each node Ni maintains a cost C(Ni,R) to every region R in the network.

 

C (Ni, R) = α distance (N, R) + (1- α) energy (N)                                          (1)

 

α=1 (pure geographic)

α=0 (pure energy)

α = 0.5 (both energy and geographic)

 

Distance = distance of the node to the centroid of the Region.

Both Distance and energy are normalized.

 

As mentioned earlier any node which has a packet to forward, asks the neighbor their respective costs to the destination region. It then forwards it to the neighbor with the minimum cost. If the minimum cost of the neighbors is more than its cost to the destination region, then it updates its cost as the minimum cost received added to one hop cost.

e.g. if a node Ni gets the minimum cost of the neighbors from node Nj to region R

 

if,  Cminimum(from neighbors, R) >  C (Ni, R) then,

            C (Ni, R) = C (Nj, R) + one_hop_cost                                                       (2)

where

One _hop_cost = α (Transmission Range*) + (1- α) (Energy spend in one transmission*)

 

* implies normalized.

 

Thus, we see that by learning the costs this way a packet will be forwarded in the path of decreasing costs.

 

2.3 Packet Termination

There are two ways in which a packet can terminate:

1. Successful Termination: When the packet reaches the destination.

2. Unsuccessful Termination:

            (A) If the destination node is dead.

(B) If the packet reaches a node which has no alive neighbors to forward the packet, except the one from which it has received the packet Figure 2 below.

 

Fig2 packet termination

 

Loop prevention is accomplished because packet is always routed in the path of decreasing minimum cost, as shown in Figure. Besides, history of last 50 packets received is maintained, and if the packet received matches with any one of them then it is ignored. In case the packet gets stuck at one node which has no neighbor, apart from the one that forwarded the packet, we drop the packet and do not retrace the path.

 

 

3 Simulation and comparison:

Following is the description of simulation tool used and simulation environment

 

3.1 Simulation tool and environment

Simulation tool used is Simjava. A simjava simulation is a collection of entities each running in its own thread. These entities are connected together by ports and can communicate with each other by sending and receiving event objects. A central system class controls all the threads, advances the simulation time, and delivers the events. In the GEAR protocol simulation each node is modeled as an entity which runs as a thread. There is a central entity which initializes the topology in a random manner. This central object which itself is also a simjava entity has the task of informing each node its location as well as its neighbor list. The progress of the simulation is recorded through trace messages produced by the entities and saved in a file.

We use a simple energy model in which every node starts with the same initial energy budget, and forwarding a packet consumes one unit of energy. Nodes initial energy level= 1 joule.

Energy depleted in transmission = 0.001 joule

So a node’s energy gets depleted in two cases:

 

We assume a MAC layer in which each node consumes negligible energy when it is not sending and receiving packets. Our routing protocol does not depend on any particular type of MAC protocol as long as it can resolve medium access control issue.

We have assumed a square area, divided into rectangular regions of fixed size. The number of nodes in each region is same.

 

3.2 Simulation Parameters

α = 0.5

Generating 50 random source and destination pairs every 5000 simulation units

 

Parameter-1

Area = 1200x1200

Area of region = 100x100

Number of Regions = 144

Number of nodes = 288

Nodes/region = 288/144 = 2

Transmission range = 130

 

3.3 Simulation Results

Below we explain the results which we obtained through our simulations. It was observed that for energy aware routing delivery fraction is approximately 1.072 times more pure geographic routing. Initially for smaller run time of simulations, both Energy aware (GEAR) and pure geographic (GPSR) give approximately identical performance. But as the simulation time increases, total numbers of nodes that die in GEAR are fewer as compared to GPSR, because we try to ensure uniform energy dissipation among the nodes in the neighborhood, by energy aware routing. Hence at higher simulation times, we observe an improvement in the performance of GEAR as compared to GPSR.

 

GEAR: Geographic and Energy Aware Routing

GPSR: Greedy Perimeter Stateless Routing (Pure Geographic)

Fig 3 Throughput vs. Simulation time

4 Conclusions

 

Geographic and Energy Aware Routing (GEAR) protocol uses energy aware and geographically informed neighbor selection to route a packet towards the target region. This strategy attempts to balance energy consumption and thereby increase network lifetime.

 

 

5 References

 

[1] Yan Yu, Ramesh Govindan and Deborah Estrin, “Geographical and Energy Aware Routing: A Recursive Data Dissemination Protocol for Wireless Sensor Networks,” UCLA Computer Science Department Technical Report UCLA/CSD-TR-01-0023, May 2001.

[2] Ad hoc networking by Charles E Perkins.

[3] David B Johnson, “Routing in ad hoc networks of mobile hosts,” in Proceedings of IEEE Workshop on Mobile Computing Systems and Applications, Dec 1994.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SCREEN SHOTS

Screen shot 1: Swing Implementation

 

Screen shot 2: Applet Implementation