![Durocher et al. proved that there is no “local” routing protocol that provides
guaranteed delivery, even under the strong assumptions of a “unit ball graph” and
accurate location information. Thus, designing a geographic routing protocol that
provides guaranteed delivery in 3-D is a challenging problem.
Existing System:
We present in this paper a novel geographic routing protocol, Multi hop
Delaunay Triangulation (MDT), that provides guaranteed delivery for a network of
nodes in a -dimensional space, for . (Only Euclidean spaces are considered in this
paper.) The guaranteed delivery property is proved for node locations specified by
arbitrary coordinates; thus, the property also holds for node locations specified by
inaccurate coordinates or accurate coordinates. We show experimentally that MDT
routing provides a routing (distance) stretch close to 1 for nodes in 2-D and 3-D
when coordinates specifying node locations are accurate. 1 When coordinates
specifying node locations are highly inaccurate, we show that MDT routing
provides a low routing (distance) stretch relative to other geographic routing
protocols. Nodes may also be arbitrarily located in a virtual space with packets
routed by MDT using the coordinates of nodes in the virtual space (instead of their
coordinates in physical space). In this case, MDT routing still provides guaranteed
delivery, but the distance stretch in physical space may be high. Geographic
routing in a virtual space is useful for networks without location information or
networks in which the routing cost between two directly connected neighbors is
neither a constant nor proportional to the physical distance between them (such as
ETT [6]). For example, a 4-D virtual space can be used for geographic routing of
nodes physically located in a 3-D space. The extra dimension makes it possible to
assign nodes to locations in the virtual space such that the Euclidean distance
between each pair of nodes in the virtual space is a good estimate of the routing
cost between them. The design of a positioning system to embed routing costs in a
virtual space is a challenging problem for wireless networks without any-to-any
routing support and beyond the scope of this paper](https://image.slidesharecdn.com/geographicroutingin-121218052100-phpapp01/85/Geographic-routing-in-2-320.jpg)
![Proposed System:
The problem is solved in a companion paper [23] where we show how to: 1)
make use Of MDT protocols to embed routing costs in virtual spaces (such as 4-
D), and 2) extend MDT routing to optimize end-to-end path costs for any additive
routing metric. MDT was designed to leverage the guaranteed delivery property of
Delaunay triangulation (DT) graphs. For nodes in 2-D, Bose and Morin proved that
greedy routing in a DT always finds a given destination node [3]. Lee and Lam
[16] generalized their result and proved that in a -dimensional Euclideanspace ,
given a destination location , greedy routing in a DT always finds a node that is
closest to . Fig. 1(a) shows a 2-D space with three large obstacles and an
arbitrary connectivity graph. Fig. 1(b) shows the DT graph [9] of the nodes in Fig.
1(a). In the DT graph, the dashed lines denote DT edges between nodes that are not
connected by physical links. The MDT graph of the connectivity graph in Fig. 1(a)
is illustrated in Fig. 1(c). By definition, the MDT graph includes every physical
link in the connectivity graph and every edge in the DT graph. In MDT routing,
when a packet is stuck at a local minimum of the connectivity graph. the packet is
next forwarded, via a “virtual link,” to the DT neighbor that is closest to the
destination. In short, the recovery method of MDT is to forward greedily in the DT
graph that is guaranteed to succeed. In this paper, we present MDT protocols for a
set of nodes to construct and maintain a correct multihop DT (formal definition in
Section II). In a multihop DT, two nodes that are neighbors in the DT graph
communicate directly if there is a physical link between them; otherwise, they
communicate via a virtual link, i.e., a multihop path provided by soft-state
forwarding tables in nodes along the path.](https://image.slidesharecdn.com/geographicroutingin-121218052100-phpapp01/85/Geographic-routing-in-3-320.jpg)
Geographic routing in
- 1. Geographic Routing in -Dimensional Spaces With Guaranteed Delivery and Low Stretch Abstract: Almost all geographic routing protocols have been designed for 2-D. We present a novel geographic routing protocol, named Multihop Delaunay Triangulation (MDT), for 2-D, 3-D, and higher dimensions with these properties: 1) guaranteed delivery for any connected graph of nodes and physical links, and 2) low routing stretch from efficient forwarding of packets out of local minima. The guaranteed delivery property holds for node locations specified by accurate, inaccurate, or arbitrary coordinates. The MDT protocol suite includes a packet forwarding protocol together with protocols for nodes to construct and maintain a distributed MDT for routing. We present the performance of MDT protocols in 3 D and 4-D as well as performance comparisons of MDT routing versus representative geographic routing protocols for nodes in 2-D and 3-D. Experimental results show that MDT provides the lowest routing stretch in the comparisons. Furthermore, MDT protocols are specially designed to handle churn, i.e., dynamic topology changes due to addition and deletion of nodes and links. Experimental results show that MDT’s routing success rate is close to 100% during churn, and node states converge quickly to a correct MDT after churn. Reasons of Proposal: Geographic routing (also known as location-based or geometric routing) is attractive because the routing state needed for greedy forwarding at each node is independent of network size. Almost all geographic routing protocols have been designed for nodes in 2-D. In reality, many wireless applications run on nodes located in 3-D. Furthermore, node location information may be highly inaccurate or simply unavailable. Consider a network represented by a connected graph of nodes and physical links (to be referred to as the connectivity graph). Greedy forwarding of a packet may be stuck at a local minimum, i.e., the packet is at a node closer to the packet’s destination than any of the nodes’s directly connected neighbors. Geographic routing protocols differ mainly in their recovery methods designed to move packets out of local minima. For general connectivity graphs in 3-D, face routing methods designed for 2-D are not applicable. Furthermore,
- 2. Durocher et al. proved that there is no “local” routing protocol that provides guaranteed delivery, even under the strong assumptions of a “unit ball graph” and accurate location information. Thus, designing a geographic routing protocol that provides guaranteed delivery in 3-D is a challenging problem. Existing System: We present in this paper a novel geographic routing protocol, Multi hop Delaunay Triangulation (MDT), that provides guaranteed delivery for a network of nodes in a -dimensional space, for . (Only Euclidean spaces are considered in this paper.) The guaranteed delivery property is proved for node locations specified by arbitrary coordinates; thus, the property also holds for node locations specified by inaccurate coordinates or accurate coordinates. We show experimentally that MDT routing provides a routing (distance) stretch close to 1 for nodes in 2-D and 3-D when coordinates specifying node locations are accurate. 1 When coordinates specifying node locations are highly inaccurate, we show that MDT routing provides a low routing (distance) stretch relative to other geographic routing protocols. Nodes may also be arbitrarily located in a virtual space with packets routed by MDT using the coordinates of nodes in the virtual space (instead of their coordinates in physical space). In this case, MDT routing still provides guaranteed delivery, but the distance stretch in physical space may be high. Geographic routing in a virtual space is useful for networks without location information or networks in which the routing cost between two directly connected neighbors is neither a constant nor proportional to the physical distance between them (such as ETT [6]). For example, a 4-D virtual space can be used for geographic routing of nodes physically located in a 3-D space. The extra dimension makes it possible to assign nodes to locations in the virtual space such that the Euclidean distance between each pair of nodes in the virtual space is a good estimate of the routing cost between them. The design of a positioning system to embed routing costs in a virtual space is a challenging problem for wireless networks without any-to-any routing support and beyond the scope of this paper
- 3. Proposed System: The problem is solved in a companion paper [23] where we show how to: 1) make use Of MDT protocols to embed routing costs in virtual spaces (such as 4- D), and 2) extend MDT routing to optimize end-to-end path costs for any additive routing metric. MDT was designed to leverage the guaranteed delivery property of Delaunay triangulation (DT) graphs. For nodes in 2-D, Bose and Morin proved that greedy routing in a DT always finds a given destination node [3]. Lee and Lam [16] generalized their result and proved that in a -dimensional Euclideanspace , given a destination location , greedy routing in a DT always finds a node that is closest to . Fig. 1(a) shows a 2-D space with three large obstacles and an arbitrary connectivity graph. Fig. 1(b) shows the DT graph [9] of the nodes in Fig. 1(a). In the DT graph, the dashed lines denote DT edges between nodes that are not connected by physical links. The MDT graph of the connectivity graph in Fig. 1(a) is illustrated in Fig. 1(c). By definition, the MDT graph includes every physical link in the connectivity graph and every edge in the DT graph. In MDT routing, when a packet is stuck at a local minimum of the connectivity graph. the packet is next forwarded, via a “virtual link,” to the DT neighbor that is closest to the destination. In short, the recovery method of MDT is to forward greedily in the DT graph that is guaranteed to succeed. In this paper, we present MDT protocols for a set of nodes to construct and maintain a correct multihop DT (formal definition in Section II). In a multihop DT, two nodes that are neighbors in the DT graph communicate directly if there is a physical link between them; otherwise, they communicate via a virtual link, i.e., a multihop path provided by soft-state forwarding tables in nodes along the path.