Efficient content caching for 5G supported vehicle networks

To evaluate the performance of the proposed scheme, a simulation environment is created against the system model. Comparative performance is analyzed by one-to-one stable matching algorithm and randomly expanded request content.

This section is divided into two subsections. The details of the simulation settings and parameters are described in the “Simulation parameters” subsection, and the comparative performance of the simulation results is discussed in the “Results and analysis” subsection.

simulation parameters

In order to compare the performance of the proposed scheme with the other two schemes, a simulation environment is created in MATLAB. In this simulation scenario, the number of RSUs with caching capacity ranges from 2 to 5, and the coverage area is randomly selected from 2 to 3 km. The cache capacity of these fog nodes ranges from 2 GB to 4 GB in increments of 1. There are vehicles randomly placed in different locations in the RSU coverage area, requesting 60 pieces of content.

The vehicles move at different speed ranges from 25 to 45 m/s. Content that is not cached in the fog nodes is retrieved from a server located at the TCC location, and the data rate for direct communication from the vehicle to the TCC is 4 MBytes/s. For communication from the vehicle to the RSU, the data rate is assumed to be 20 MBytes/s.^{twenty four}.

A complete list of simulation parameters is shown in Table 1. To fairly evaluate the performance of the proposed scheme, Monte Carlo-based simulations with other schemes are performed and the results are obtained as the average of the following schemes: $10^5$ experiment.

Table 1 Simulation parameters.

Results and analysis

The performance of the proposed scheme is evaluated in terms of the percentage of content cached in RSU, cumulative and average content download time, and downloaded data. Simulation results are obtained for different numbers of RSUs, different content cache capacities, and different numbers of incoming content requests.

Popularity of cached content

The popularity of cached content on an RSU is calculated by determining the ratio of the total number of content cached on all fog nodes to the total number of requested content received by the TCC.

The results shown in Figure 3 represent the percentage of total content cached in fog-enabled RSUs as the number of RSUs increases. The results show that the proposed scheme cached 25% of the requested content when the number of RSUs was only two, which increases to about 38% for five RSUs with the same cache capacity. I am. However, the content cached in the other two schemes is much less than that proposed for various numbers of RSUs. The same trend is shown in Figure 4 when the content cache capacity of the three fog-enabled RSUs increases from 2 GB to 4 GB of data. The results show that the number of cached contents increases as the RSU cache capacity increases from 42% to 67%. However, the content cache capacity of 1:1 and FCFS is much smaller than the proposed scheme.

Figure 5 shows the percentage of popular content cached in RSU for different numbers of content requests. This result shows that content is cached in the fog-enabled RSU when the content requests are less than the maximum value. However, as the number of content requests increases, the percentage of content cached in the RSU gradually decreases. From these results, it is clear that the proposed scheme is still able to cache a high percentage of popular content in his RSUs compared to the other two schemes.

Content download time

The content download time for a content requesting vehicle is calculated as the time when the vehicle sends the content request and the entire content is received. In this work, we considered three possible scenarios in which a vehicle retrieves the required content.

1.

If the vehicle needs to retrieve content from a server located in a remote location, this information is retrieved at a lower data rate (DR1) As mentioned in Equation 1, the download time will be longer. (Ten).
2.

When a vehicle’s requested content is located on a connected fog node-based RSU, the requested content is significantly shorter because both the vehicle and RSU are on the same network and provide high data rates. Downloaded in hours (DR2). The download time from connected fog nodes is calculated as described in Equation 1. (11).
3.

If the vehicle cannot download the entire content from the fog node due to the short connection time with the fog node, it must retrieve some of the data from a remotely located server. Suppose a node needs to fetch. CS Contents and parts of them ($CS_1$) is downloaded from the connected fog nodes, and the remaining content is downloaded from a remotely located server, increasing the total download time (DT) is calculated as follows.

$$\begin{Alignment} DT = \frac{CS_1}{DR1} + \frac{CS-CS_1}{DR2} \end{Alignment}$$

The results are shown in FIGS. 2 to 4. Figures 6 and 7 calculate the average time it takes for all moving requesting vehicles to download a certain number of contents. Figure 6 shows that the average download time to download all requested content for different numbers of RSUs is shorter than the other two of his schemes. As the number of RSUs increases, the download time continues to decrease. This is because these RSUs have a large amount of cached content, and most of the requested content is downloaded at a higher data rate and in a shorter amount of time. The same trend is seen in Figure 7. Here, the average download time for the same number of requesting nodes is calculated for different cache sizes of fog-enabled RSUs. From the results, it is clear that the proposed scheme is able to download the requested content in a relatively short time compared to the other two schemes. Additionally, increasing the cache capacity of fog nodes reduces the average download time.

Figure 8 shows the average content download time from different numbers of content requests from vehicles when the number of RSUs and their cache capacity remain the same. The results show that for all three schemes, the average download time to retrieve all requested content increases with the increase in content requests. However, in the proposed scheme, this download time is significantly shorter than the other two of his schemes.

Figures 9 and 10 show the calculated total time for a vehicle to download all requested content for a fixed number of content requests and for varying the number of RSUs and cache capacity, respectively. I am. As shown in Figure 9, the content download time continues to decrease as the number of RSUs increases because the RSUs cache most of the requested content, resulting in faster download times. The same trend continues with changes in RSU content cache capacity. As shown in Figure 10, we request the same number of requested contents with a fixed number of RSUs and their cache capacity. These results show that the proposed scheme has significantly lower download time for all requested content compared to his other two schemes.

The results shown in Figure 11 calculate the total time required to download all the requested content from the cached remote server. From the results, it is clear that the total download time of the proposed scheme is shorter than the other two of his schemes. However, this time will continue to increase as the content request time increases over a given period of time.

Downloaded data

Download data for all content cached on fog-enabled RSUs is calculated in this section. The results are shown in FIGS. 2 and 3. From FIGS. 12 and 13, it can be seen that the total download data amount of content cached from RSU in the proposed method is larger than in the other two methods. This is due to the maximum number of popular content cached in RSU.

Figure 12 shows that to increase the number of RSUs, the data downloaded from cached content in the proposed scheme increases from 8 GB to 28 GB for 2 to 5 RSUs, respectively. However, the other two schemes download less data from cached content. The same trend is shown in Figure 13. If the fog node cache size increases for a fixed number of RSUs, i.e. 3. The results show that when the cache capacity of each RSU is 4, the proposed scheme downloads 24 GB of data. G.B. However, the other two schemes only allowed you to download 16 GB of data.

In the results shown in Figure 14, we observe that data is downloaded from cached content for different numbers of content requests received from 100 vehicles. From the results, it is clear that the proposed scheme downloads more data from cached content compared to his other two schemes. As the number of content increases, less data is downloaded from cached content. This is because fog nodes cannot cache all requested content to a fixed number of RSUs, resulting in less data being downloaded from the same number of vehicles.