After describing the design concept of the proposed 3D-GAR scheme, the implementation issues of the proposed protocol consisting of both the GF and the 3D-RUT algorithms are explained in this section. The GF scheme is considered a sequential table-lookup algorithm that only requires the implementation of the one-hop neighbor table. Therefore, both the time and space complexities are O(m), where m represents the number of neighbors specified in the one-hop neighbor table. If the void problem occurs, the 3D-RUT scheme is utilized to forward packets to the nodes in the boundary node set B as defined in Definition 11.
Since a node Ni’s neighbors in the boundary node set can construct rolling balls with Ni, the original mechanism of forwarding packets to the nodes in B can therefore be transformed into a simple forwarding rule. In other words, node Ni which currently conducts the 3D-RUT scheme simply forwards packets to those neighbors that can form a rolling ball with Ni, where these neighbors can be obtained by the following method. For each pair of nodes (Nj, Nk) in the one-hop neighbor table of Ni, if a node-free-inside circumscribed ball hinged at Ni with a radius of R/2 can be established with both nodes Nj and Nk on the surface of the ball, the definition of the rolling ball (i.e., Definition 9) will be satisfied since the two conditions are satisfied as follows: (a) node Ni is located on the surface of the ball; and (b) there does not exist any neighbor node situated inside the ball. As a result, both Nj and Nk are considered as the next hopping nodes of Ni for packet forwarding. It is noted that the time complexity of this process is O(m3) since three nested loops to go through the one-hop neighbor table are required to conduct this procedure; while the space complexity is still O(m) since only the construction of the neighbor table is required. Finally, by considering both the GF and the 3D-RUT schemes, the time and space complexities of the 3D-GAR protocol can be acquired as O(m3) and O(m) respectively, where m is the number of neighbors specified in the one-hop neighbor table.
400 450 500 550 600 650 700 750 800
400 450 500 550 600 650 700 750 800
5
(b) Path Length for Successful Transmission
400 450 500 550 600 650 700 750 800
101
Figure 3.3: Performance evaluation for the proposed 3D-GAR protocol.
3.5 Performance Evaluation
The performance of the proposed 3D-GAR algorithm is evaluated and compared with other three protocols via simulations, including the 3D-ABLAR, the network flooding, and the reference GF algorithms. The simulation settings are explained as follows. A number of 1000 SNs are randomly deployed in the Euclidean 3D box ranging from (0, 0, 0) to (1000, 800, 1200) in the unit of meters. The transmission range of a node is 250 m. A pair of source and destination nodes are respectively located at (0, 400, 600) and (1000, 400, 600). The source node is with the data transmitting rate of 16 Kbps and data packet size of 512 bytes. There also exists a void block with length 400 m, width 800 m, and variable heights. This void block is randomly placed in the network in order to simulate the occurrence of void problems. In other words, there are SNs around the peripheral of the void block; while none of the nodes is situated inside the void block. Three performance metrics are utilized in the simulations for performance comparison as follows: (a) packet arrival rate: the ratio of the number of received data packets to the number of total data packets sent by the source; (b) path length: the average path length of successful routing in the unit of hop count; and (c) routing overhead:
the average number of transmitted bytes per second for a network node.
Fig. 3.3(a) shows the packet arrival rate performance versus the void height. Due to the property of guaranteed delivery, both the 3D-GAR and network flooding algorithms will result in the delivery rate of 100%. The 3D-ABLAR and the GF protocols incur less delivery rate with regard to the increase of void height since the void problem occurs frequently when the void height becomes large. The 3D-ABLAR protocol has higher delivery rate than the GF algorithm owing to the reason that the GF scheme drops packet directly as the void problem occurs. Fig. 3.3(b) illustrates the performance of average path length for successful routing versus the void height. The network flooding algorithm results in the lowest value under small void heights since it can always find the shortest path between the source and the destination.
However, as the void height becomes large, both the 3D-ABLAR and GF schemes will incur relatively smaller value of path length. The major reason is due to the 100% of packet delivery rate from the network flooding algorithm, which requires additional packet rerouting as the
void problem becomes severe. The same reason can be applied to the curve obtained from the 3D-GAR protocol, which possesses a slightly longer routing path than the other protocols, e.g., around 1.8 additional hops under the void height of 800 m. Nevertheless, even with the slightly larger path length, the 3D-GAR protocol can result in guaranteed delivery rate as shown in Fig. 3.3(a), which outperforms both the 3D-ABLAR and GF schemes with lowered packet delivery rate.
Fig. 3.3(c) shows the performance of routing overhead versus the void height. It can be observed that both the 3D-ABLAR and GF schemes result in comparatively low overhead since most of the packets are dropped due to the void problem. In order to guarantee delivery and to find the shortest path, the network flooding algorithm generates a large number of packets in comparison with other protocols, which contributes to a significant amount of routing overhead as shown in Fig. 3.3(c). Comparing with the network flooding algorithm, the 3D-GAR protocol can achieve guaranteed packet delivery with a comparably smaller number of packets since the 3D-GAR scheme limits the packet rerouting only to nodes that are in the boundary node set. The merits of the proposed 3D-GAR protocol can therefore be observed, which achieves guaranteed packet delivery with reasonable routing overhead.
3.6 Summary
In this chapter, a three-dimensional greedy anti-void routing (3D-GAR) protocol is proposed to completely resolve the void problem incurred by the conventional greedy forwarding algo-rithm under the 3D environment. The 3D rolling-ball UBG boundary traversal (3D-RUT) scheme is adopted within the 3D-GAR protocol to solve the boundary finding problem, which results in the guarantee of packet delivery. In the end, the correctness proofs, protocol im-plementation, and performance evaluation of the proposed algorithms are properly provided.
According to the guaranteed packet delivery and the relatively low routing overhead proper-ties, the proposed 3D-GAR protocol can be utilized to implement the green wireless access networks as a unicast routing protocol in the three dimensional space.
Chapter 4
Component-based Routing Platform and Energy Conserving Multicast Routing Protocol
Chapter Overview
In the network layer multicast protocol design for achieving the green wireless access networks, how to provide low energy consumption and maintain high packet delivery ratio is considered the major issue. Similar to the previous chapters, the main theme for this chapter is chosen as in the wireless sensor network (WSN) due to its stringent requirements on the power consumption. With regard to the multicast protocol design, reducing the number of data transmissions is considered a feasible method for decreasing the energy consumption of the network nodes. In the wired networks, the Steiner-tree is regarded as the optimal approach for constructing the multicast structure for specific senders and receivers. However, the results can not be directly applied to the wireless environment. Furthermore, in the aspect of protocol realization, the conventional implementations of routing algorithms are designed to be protocol specific, which results in the lack of flexibility for developers to implement other protocols.
In this chapter, an energy conserving multicast routing (ECMR) protocol is proposed to
reduce the total number of relaying nodes for the construction of a multicast tree. It is designed to be a heuristic scheme since achieving a minimal cost multicast tree is considered an NP-hard problem in the wireless broadcast environment. Moreover, a component-based routing platform (CRP) with encapsulated software components (ESCs) is proposed to provide a generic routing implementation platform. Based on the design of the proposed ESCs, routing algorithms can be implemented in a more efficient manner. The ECMR algorithm can be implemented on the proposed CRP practical embedded platforms for performance evaluation. It validates that the CRP can be adopted as an effective design platform for the implementation of routing algorithms. Compared with the existing multicast routing protocol in the field experiments, the experimental results show that the proposed ECMR scheme can provide better energy conservation while the packet delivery ratio is still preserved. Thanks to the design of low energy consumption with high delivery ratio, the proposed ECMR protocol can be utilized in the establishment of the green wireless access networks as a feasible multicast routing protocol.
4.1 Introduction
Wireless sensor networks (WSNs) are composed by numerous sensor nodes (SNs) with in-formation processing, sensing, and wireless communication capabilities. Various tasks can be cooperatively performed by the SNs, including monitoring, locationing, and data distri-bution [40]. Due to the distributed nature and hardware cost associated with the SNs, the available resources are considered limited for most of the current applications. Severe power constraint is considered one of the most challengeable design issues due to the difficulty in battery recharging for a large amount of SNs, especially in a remote or dangerous region.
The battery lifetime within an SN can be extended by technological advancement under different fields, e.g., hardware, software, or network design. Appropriate hardware design can fulfill the performance requirements that are agreeable to the least power consumption [41].
Dynamic power management within the software layers, e.g., application software or operating system, can also achieve energy conservation by adaptively adjusting the power modes of an
SN [42]. Moreover, energy conservation algorithms can be applied at the network level with the cooperation between the SNs. By adopting appropriate design of routing algorithms [43, 44]
for packet delivery, the lifetime of the entire WSN can therefore be prolonged.
With the emergent applications for multicast data dissemination and aggregation, the design of multicast routing algorithms are considered one of the important topics within the WSNs. Different types of multicast routing protocols have been proposed for the mobile ad hoc network (MANET). Basically, these protocols can be categorized into the tree-based (such as MAODV [45], AMRIS [46], AMRoute [47], and GS [48]) and the mesh-based (such as ODMRP [49], CAMP [50], and MCEDAR [51]) algorithms. The fundamental topology of the tree-based network originates from a root, which stretches out its branches and the corresponding subbranches. The benefit of using the tree-based structure is its low energy consumption and the simplicity for maintaining the structure in static networks. In the mesh-based structure, there are at least two routes between each receiver and the transmitter. Compared with the tree-based networks, this architecture can provide more robust connectivity under the channel fading effect. However, the inherent complexity within the mesh-based structure makes it consume more system resources. Most of the existing multicast routing protocols designed for the MANET can not be directly adopted within the WSN for its stringent energy requirements. Moreover, these routing algorithms are considered infeasible for practical implementation either due to the protocol complexity or the excessive rerouting expenses.
Moreover, in terms of evaluating the routing algorithms, most of the existing research work conducts computer simulations for performance comparison. However, computer sim-ulations are in general considered impractical due to the assumption of ideal network con-figurations. Therefore, field experiments with implementation testbeds become necessary for the evaluation of protocol performance. There are existing works that implement the ad hoc routing algorithms with field experiments, including Kernel-AODV [52], AODV-UCSB [53], and AODV-UU [54] schemes. However, these implementations are specifically designed to facilitate the design of their own routing protocols. It will be beneficial to provide a generic platform associated with development toolkits for evaluating different routing algorithms.
In this chapter, an energy conserving multicast routing (ECMR) protocol is proposed
for reducing the number of data transmissions within the WSNs. Since obtaining the mini-mal cost multicast tree under the wireless broadcast environment has been proven to be an NP-hard problem [55], a heuristic algorithm is exploited by the ECMR scheme for energy conservation purpose. By adopting the proposed algorithm, the conventional multicast struc-ture (e.g., the ODMRP protocol [49]) can be converted into a light-weight topology. The total energy consumption can therefore be decreased; while the packet delivery ratio is still maintained. Furthermore, a component-based routing platform (CRP) is also proposed as a generic implementation testbed based on the Linux embedded system for the protocol realiza-tion. Field experiments are conducted in order to evaluate the performance of the proposed ECMR algorithm and validate the proposed CRP implementation platform. As can be seen from the experiment results, the proposed ECMR protocol is capable of providing better energy conservation comparing with the existing ODMRP algorithm.
The remainder of this chapter is organized as follows. The proposed ECMR protocol is presented in Section 4.2. Section 4.3 describes the proposed CRP implementation platform, including the software and hardware systems. Based on the proposed CRP testbed, the field experiments of the proposed ECMR protocol are conducted and compared with the baseline