CN102724731B - Epidemic routing algorithm with adaptive capacity - Google Patents

Abstract

The invention relates to an opportunistic network routing algorithm, which is used to improve the Epidemic routing algorithm so that nodes in the opportunistic network can efficiently forward data packets while reducing network resource consumption as much as possible. The Epidemic routing algorithm can achieve high transmission success rate and low transmission delay in some scenarios, but the adaptability of the algorithm is poor, and in other scenarios, the performance of the algorithm will drop sharply. The invention proposes an adaptive mechanism, and uses the mechanism to improve the Epidemic routing algorithm. The adaptive mechanism can effectively reduce the number of copies of invalid data packets in the network, reduce the consumption of network resources, improve the performance of the routing algorithm, and then improve the scalability of the Epidemic routing algorithm.

Description

Epidemic Routing Algorithm with Adaptive Capability

technical field

The invention relates to an opportunistic network routing algorithm. The function is to enable nodes in the opportunistic network to efficiently forward data packets, and at the same time reduce the forwarding amount of nodes as much as possible, thereby reducing network resource consumption.

Background technique

An opportunistic network is a self-organizing network that can tolerate network communication, delay and split by using the encounter opportunities brought about by node movement without the need for a complete path between the source node and the destination node. Opportunistic networks are different from traditional multi-hop wireless networks in that their nodes are not uniformly deployed, the network scale and the initial positions of nodes are not preset, and the path between source nodes and destination nodes cannot be determined in advance. Opportunistic networks use the "store-carry-forward" mode to transmit information hop by hop to achieve inter-node communication. Its architecture is different from multi-hop wireless networks. It inserts a new protocol called bundle layer between the application layer and the transport layer. layer.

Because opportunistic networks can deal with problems that are difficult to solve with traditional wireless network technologies such as network splitting and delay, and can meet the needs of network communication under harsh conditions, it is mainly used in situations where there is a lack of communication infrastructure, harsh network environments, and emergencies. .

1. Contrast routing algorithm

In order to compare with the routing algorithm of the present invention, two typical routing algorithms are selected as reference algorithms. Epidemic algorithm is a typical representative of routing algorithm based on flooding strategy, and many routing algorithms based on flooding strategy can be regarded as derived from this algorithm. The Spray and Wait algorithm performs flooding according to a certain strategy and is based on a limited flooding strategy. The main performance indicators of the algorithm have significant advantages in most scenarios.

(1) Epidemic algorithm

The basic idea of the Epidemic algorithm is that when two nodes meet, they exchange data packets that the other party does not have. After enough exchanges, theoretically each non-isolated node will receive all data packets, thereby realizing the transmission of data packets.

In the Epidemic algorithm, each data packet has a globally unique identifier, and a summary vector is saved in each node to record the data packets carried in the node. When two nodes meet, the two parties first exchange summary vectors, and after knowing the data packets carried by the other party, the two parties only transmit the data packets that the other party does not have.

The Epidemic algorithm is essentially a flooding algorithm. In theory, this algorithm can maximize the success rate of data packet transmission and minimize transmission delay, but it will also cause a large number of data packet copies in the network and consume a large amount of network resources. .

The Epidemic algorithm has three goals, which are the maximum transmission success rate, the minimum transmission delay and the minimum network resource consumption. Achieving the above goals requires specific scenarios. In most scenarios, the performance of routing algorithms is significantly degraded due to excessive flooding.

(2)Spray And Wait algorithm

The Spray and Wait algorithm is divided into two stages. The first is the Spray stage, in which some data packets in the source node are diffused to neighboring nodes; then it enters the Wait stage, if the target node is not found in the Spray stage, the node containing the data packet transmits the data packet to the target node in Direct Delivery mode, that is Only when the target node is encountered, the data packet is sent. The transmission volume of this algorithm is significantly less than that of the Epidemic algorithm, the transmission success rate is high, the transmission delay is small, and the algorithm has strong applicability.

(3) Prophet algorithm

The Prophet algorithm is based on a probability strategy. This routing algorithm estimates the probability of successful packet transmission, selectively copies data packets, and tries to avoid generating copies with low transmission efficiency. The algorithm defines a transmission prediction value to describe the probability of successful transmission between nodes. When two nodes meet, the nodes update their respective transmission prediction values, and use this value to decide whether to forward the data packet.

2. Metrics

The metrics for evaluating the performance indicators of opportunistic network routing algorithms mainly include:

(1) Transmission success rate

Delivery Ratio is the ratio of the total number of data packets that successfully arrive at the target node within a certain period of time and the total number of data packets that need to be transmitted from the source node. This indicator describes the ability of the routing algorithm to correctly forward data packets to the target node. the most important indicator.

(2) Transmission delay

Delivery Delay (Delivery Delay) is the time required for a data packet to reach the destination node from the source node, and is usually evaluated by the average delivery delay. Small transmission delay means that the routing algorithm has strong transmission capability and high transmission efficiency, and it also means that less network resources will be occupied during the transmission process.

(3) Routing overhead

Routing overhead (Overhead) refers to the total number of data packets forwarded by nodes within a certain period of time, and is usually evaluated by the ratio of the number of data packets successfully reaching the target node to the total number of data packets forwarded by all nodes. High routing overhead means that nodes forward a large number of data packets, which will flood the network with a large number of data packet copies, increase the probability of data packet collisions, and consume a large amount of node energy.

3. Performance analysis of Epidemic algorithm

Based on the scenario in Table 1, the total number of data packets is 50 and each node generates 10 data packets for simulation respectively, and the results shown in Figure 1 and Figure 2 are obtained.

In Figure 1 and Figure 2, Spray And Wait is used as a comparison algorithm. This algorithm can obtain near-optimal transmission success rate and routing overhead in most scenarios, and can maintain good performance regardless of the size of the network. scalability.

From Figure 1 and Figure 2, the following conclusions can be drawn:

(1) In some specific scenarios, the Epidemic algorithm has a very high transmission success rate and a very low transmission delay, which is much better than the comparison algorithm in these two indicators;

(2) When the number of data packets is constant, the increase in the number of nodes in the network will improve the performance of the routing algorithm;

(3) In some scenarios, there are some factors closely related to the network application environment that will lead to a significant decline in the performance of the Epidemic algorithm.

Based on the scenario in Table 1, Figure 3 describes the relationship between the number of data packets and the success rate of transmission when the total number of nodes is certain. It can be seen from Figure 3 that when the number of data packets increases, the success rate of transmission decreases accordingly. The present invention refers to the cause of this phenomenon as the crowding out effect, that is, when the total number of data packets to be transmitted in the network exceeds the total amount of data packets that can be stored by the node, the node cache saturation phenomenon will occur, and the node receives new data at this time When the packets are processed, the old data packets have to be discarded according to certain rules. The existence of this effect leads to a significant decline in the performance of the Epidemic algorithm.

Contents of the invention

The invention relates to a new opportunistic network routing algorithm, which introduces an adaptive mechanism on the basis of the Epidemic routing algorithm. The algorithm can reduce the amount of forwarding of invalid data packet copies, obtain higher transmission success rate and lower consumption of network resources.

The crowding out effect in the Epidemic algorithm is the core reason for the performance degradation of the algorithm. Reducing the number of copies of data packets in the network can suppress the crowding out effect, but if the number of copies is too small, the performance of the algorithm will also decline. If the number of data packet copies sent can be determined according to the current status of node caches in the network, and a better compromise can be achieved, the performance of the algorithm can obviously be improved and the applicability of the algorithm can be expanded.

The invention improves the Epidemic algorithm, and aims to actively reduce the number of copies of data packets injected into the network when the cache of nodes in the network tends to be full, and suppress the occurrence of crowding out effect, even though the algorithm has self-adaptive ability. Concrete scheme is to increase following mechanism on the basis of Epidemic algorithm, the present invention calls it adaptive mechanism, and adopts this mechanism algorithm to be called Adaptive Epidemic algorithm.

(1) Each node maintains a field to store the threshold λ, λ∈[0,1]. If the percentage of occupied space in the node cache exceeds the threshold λ, the node cache is considered saturated;

(2) When node i meets any node j, node i first obtains the cache status of j and surrounding nodes, counts the number N _ai of nodes in contact with node i, and the number N _ei of buffer saturated nodes;

(3) Calculate p _i =N _ei /N _ai , from its definition we can know that p _i ∈ [0,1];

(4) Node i randomly copies the data packet according to the value of p _i and sends it to the node in contact with it.

Under the self-adaptive mechanism, the p value can reflect the cache saturation of surrounding nodes, and injecting data packet copies into the network according to the p value can obviously suppress the common occurrence of cache saturation and the crowding out effect, thereby improving the performance of the routing algorithm.

Description of drawings

Figure 1 Comparison of transmission success rate

Figure 2 Transmission delay comparison

Figure 3 The impact of the number of data packets on the transmission success rate

Figure 4 The influence of the threshold value on the transmission success rate

Figure 5 The transmission success rate of the improved algorithm in different scenarios

Figure 6 The transmission delay of the improved algorithm in different scenarios

Figure 7 Routing overhead of the improved algorithm in different scenarios

Detailed ways

The principles and features of the present invention are described below, and the examples given are only used to explain the present invention, and are not intended to limit the scope of the present invention.

Use ONE (the Opportunistic Networking Environment) simulation platform to implement the routing algorithm involved in the present invention. In the following simulation, pedestrians carrying smart Bluetooth devices are simulated walking in real urban scenes, and used to implement and analyze the performance of routing algorithms. The specific scene settings are shown in Table 1.

Table 1 Simulation scene settings

(1) The influence of the threshold

Based on the scenario in Table 1, take 100 nodes and 800 data packets as an example to analyze the influence of the threshold on the success rate of algorithm transmission. The results are shown in Figure 4. The data in Figure 4 is calculated by (d _ae -d _e )/d _e , where d _ae and d _e are the transmission success rates of AdaptiveEpidemic and Epidemic algorithms, respectively.

According to the Adaptive Epidemic algorithm mechanism, when the threshold value is 0, the algorithm degenerates into the Epidemic algorithm, and the results in Figure 4 also verify this conclusion. It can be seen from Figure 4 that when the threshold is set properly, the transmission success rate of the algorithm can be greatly improved, for example, when the threshold is 90%, the improvement reaches 38.9%.

(2) The impact of the number of different data packets

Based on the scenario in Table 1, use 100 nodes, 30-3600 data packets, and a data packet lifetime of 3 hours to evaluate the performance of the algorithm.

In Figure 5-Figure 7, Epidemic, Prophet, Spray And Wait are used as comparison algorithms. Prophet and AdaptiveEpidemic belong to the same kind of algorithms, and both can be regarded as improving the performance of the algorithm by limiting the number of data packet copies on the basis of the Epidemic algorithm. The two algorithms are highly comparable. The reason for comparison with the Spray And Wait algorithm is that the algorithm has strong adaptability and can achieve better performance in various network environments.

When the number of data packets is 30, the node cache will not be saturated due to the small number of data packets. At this time, the Adaptive Epidemic algorithm degenerates into the Epidemic algorithm. The simulation results in Fig. 5-Fig. 7 show that all the indicators of the 2 algorithms are the same, which is consistent with the theoretical analysis results.

The results in Figure 5 show that the transmission success rate of the Adaptive Epidemic algorithm is better than that of Epidemic and Prophet under different numbers of data packets. For example, when there are 1600 data packets, the degree of improvement reaches 39.3% and 25.6% respectively.

The results in Figure 5 show that, compared with the Spray And Wait algorithm, Adaptive Epidemic can take full advantage of Epidemic when the number of data packets is small, and the transmission success rate is significantly ahead. When there are many data packets, Adaptive Epidemic can also achieve performance better than or close to the Spray And Wait algorithm because the adaptive mechanism suppresses the transmission of data packet copies.

The results in Figure 6 show that the transmission delay index of the Adaptive Epidemic algorithm is worse than that of the Epidemic algorithm. The largest gap occurs when 800 data packets are received, and the transmission delay increases by 17.6%. This result is in line with the algorithm mechanism. When the node buffer area is saturated , the packet needs to wait in the source node, which increases the transmission delay.

The results in Figure 7 show that the Adaptive Epidemic algorithm reduces the overall routing overhead compared with the Epidemic algorithm, especially when there are many nodes, such as 3600 nodes, the decrease is 87.1%; compared with the Spray And Wait algorithm, it is also decreased by 79.2%

It can be seen from Figures 5 to 7 that the trends of the indicators of Adaptive Epidemic, Epidemic and Prophet algorithms are highly consistent, reflecting the internal connection of the three algorithms; it also shows that the self-adaptive mechanism will not change the core mechanism of the Epidemic algorithm.

The simulation results in Figures 5-7 support the previous theoretical analysis, showing that in some scenarios, the adaptive mechanism can suppress the crowding out effect to a certain extent, improve the performance of the algorithm, and broaden the applicability of the algorithm.

Claims (1)

1. an opportunistic network routing algorithm, it is characterized in that: this routing algorithm comprises the principle of Epidemic routing algorithm, parameter and the course of work, that the one of Epidemic routing algorithm is improved, introduce adaptation mechanism on the basis of Epidemic routing algorithm, in opportunistic network, each node maintenance field is used for depositing threshold values λ, λ ∈ [0,1], if the percentage in occupied space exceedes threshold values λ in nodal cache, then think that nodal cache district is saturated; When node i and any node j meet, first node i obtains node j and surroundings nodes cache condition, adds up the node number N contacted with node i _ai, cushion saturated node number N _ei; Calculate p _i=N _ei/ N _ai, define known p by it _i∈ [0,1]; Node i is according to p _ivalue, random reproduction packet is also sent to the node contacted with it.