Academia.eduAcademia.edu

Figure 9 – uploaded by Tahar Allaoui

See full PDFdownloadDownload figure
Figure 10 Average waiting time to enter the CS of NFK compared to CKM with respect to the number of requests (see online version for colours) Figure 11 Amount of messages exchanged of NFK compared to CKM with respect to the communications range (see online version for colours)

Figure 10 Average waiting time to enter the CS of NFK compared to CKM with respect to the number of requests (see online version for colours) Figure 11 Amount of messages exchanged of NFK compared to CKM with respect to the communications range (see online version for colours)

Related Figures (20)

The final publication is available at
The final publication is available at
Algorithm 3 Reception of request (Token, j) Following Algorithm 4, when a requesting site receives a reject message, it is going tc look for another adequate site holding a free token by consulting its routing table. Using Algorithm 3, if a site receives a request message, it is going to send the token towards the requesting site if it does not use the resource. Otherwise, it puts the requesting site on hold. In case the site does not hold a token any more, it is going to send a reject message to the requesting site. Algorithm 4 Evolutionary algorithm
Algorithm 3 Reception of request (Token, j) Following Algorithm 4, when a requesting site receives a reject message, it is going tc look for another adequate site holding a free token by consulting its routing table. Using Algorithm 3, if a site receives a request message, it is going to send the token towards the requesting site if it does not use the resource. Otherwise, it puts the requesting site on hold. In case the site does not hold a token any more, it is going to send a reject message to the requesting site. Algorithm 4 Evolutionary algorithm
Algorithm 5 Evolutionary algorithm In Algorithm 6, the modification in the routing table concerns two variables: the existence of a free token, and the existence of a present token, according to the values of X and Y.
Algorithm 5 Evolutionary algorithm In Algorithm 6, the modification in the routing table concerns two variables: the existence of a free token, and the existence of a present token, according to the values of X and Y.
Algorithm 7 Find the adequate token holder(AODVi) Procedure 7 presents the judicious manner that we use to find the closest token holder for the requesting site.
Algorithm 7 Find the adequate token holder(AODVi) Procedure 7 presents the judicious manner that we use to find the closest token holder for the requesting site.
5.2 Fault situations ure 2 Proposed message format extension related to the fault-tolerance (see online version for colours) We can find several cases of faults in our system, where the fault can occur at the level of:
5.2 Fault situations ure 2 Proposed message format extension related to the fault-tolerance (see online version for colours) We can find several cases of faults in our system, where the fault can occur at the level of:
In this case, the back-to-back site becomes a safe site, and it chooses a new site to become its back-to-back site. The latter will receive a copy of the waiting queue from its safe site, as illustrated in Figure 4. Figure 3 illustrates the tolerance strategy when the fault occurs at the token holder level.
In this case, the back-to-back site becomes a safe site, and it chooses a new site to become its back-to-back site. The latter will receive a copy of the waiting queue from its safe site, as illustrated in Figure 4. Figure 3 illustrates the tolerance strategy when the fault occurs at the token holder level.
Figure 5 Tolerance strategy when the fault occurs at the back-to-back site level (see online version for colours)
Figure 5 Tolerance strategy when the fault occurs at the back-to-back site level (see online version for colours)
Figure 6 illustrates the tolerance strategy when the fault occurs at both the token holder and its safe site at the same time.
Figure 6 illustrates the tolerance strategy when the fault occurs at both the token holder and its safe site at the same time.
Figure 7 illustrates the tolerance strategy when the fault occurs at both a site holding a token and a back-to-back site at the same time.
Figure 7 illustrates the tolerance strategy when the fault occurs at both a site holding a token and a back-to-back site at the same time.
ure 8 Tolerance strategy when the fault occurs simultaneously at a safe site and its back-to-back site (see online version for colours) Figure 8 illustrates the tolerance strategy when the fault occurs at both a safe site and its back-to-back site at the same time.
ure 8 Tolerance strategy when the fault occurs simultaneously at a safe site and its back-to-back site (see online version for colours) Figure 8 illustrates the tolerance strategy when the fault occurs at both a safe site and its back-to-back site at the same time.
Table 2 Simulation parameters We evaluated the performance of our solution on different scenarios by varying: number of requests, network density, communications range, and number of faults. In addition, for every scenario, we studied the amount of exchanged messages together with the generated delay. We also compared the waiting time and number exchanged messages of our proposal NFK to those obtained by CKM (Haghighat and Mohamadi, 2011), we chose to compare with CKM mainly because it was evaluated within the same environment as ours, and using the same routing strategy.
Table 2 Simulation parameters We evaluated the performance of our solution on different scenarios by varying: number of requests, network density, communications range, and number of faults. In addition, for every scenario, we studied the amount of exchanged messages together with the generated delay. We also compared the waiting time and number exchanged messages of our proposal NFK to those obtained by CKM (Haghighat and Mohamadi, 2011), we chose to compare with CKM mainly because it was evaluated within the same environment as ours, and using the same routing strategy.
‘igure 9 Amount of exchanged messages with respect to the number of requests (see online version for colours) Figure 9 presents the amount of exchanged messages with respect to the number of requests both with and without fault-tolerance mechanisms. It is clear that the average number of messages is much higher for the algorithm with fault tolerance because of the different updates required to implement the required tolerance. However, this double-layer system achieved a similar delay despite the presence of faults, as illustrated in Figure 10.
‘igure 9 Amount of exchanged messages with respect to the number of requests (see online version for colours) Figure 9 presents the amount of exchanged messages with respect to the number of requests both with and without fault-tolerance mechanisms. It is clear that the average number of messages is much higher for the algorithm with fault tolerance because of the different updates required to implement the required tolerance. However, this double-layer system achieved a similar delay despite the presence of faults, as illustrated in Figure 10.
figure 12 Average waiting time to enter the CS with respect to the communications range (see online version for colours) Figure 13 Amount of exchanged messages with respect to network density (see online version for colours)
figure 12 Average waiting time to enter the CS with respect to the communications range (see online version for colours) Figure 13 Amount of exchanged messages with respect to network density (see online version for colours)
‘igure 14 Average waiting time to enter the CS with respect to network density (see onlin¢ version for colours) Figure 15 Amount of exchanged messages with respect to the number of ressources (see online version for colours)
‘igure 14 Average waiting time to enter the CS with respect to network density (see onlin¢ version for colours) Figure 15 Amount of exchanged messages with respect to the number of ressources (see online version for colours)
gure 16 Average waiting time to enter the CS with respect of the number of resources (see online version for colours) ‘igure 17 Amount of exchanged messages in respect of the number of faults (see online version for colours)
gure 16 Average waiting time to enter the CS with respect of the number of resources (see online version for colours) ‘igure 17 Amount of exchanged messages in respect of the number of faults (see online version for colours)
From both Figures 17 and 18, we can see that the number of messages and the waiting time increase logically when increasing the number of faults due to the different fault-tolerance updates. However, our proposal is able to sustain low delays (<0.4 seconds in the worst case).
From both Figures 17 and 18, we can see that the number of messages and the waiting time increase logically when increasing the number of faults due to the different fault-tolerance updates. However, our proposal is able to sustain low delays (<0.4 seconds in the worst case).
Figure 18 Average waiting time to enter the CS with respect to the number of faults (see online version for colours) 7 Conclusions and future work
Figure 18 Average waiting time to enter the CS with respect to the number of faults (see online version for colours) 7 Conclusions and future work
Related topics:
Computer ScienceDistributed ComputingMobile Ad Hoc NetworksInformation Communication TechnologyMobile Ad Hoc NetworkMutual ExclusionWireless Ad Hoc Networkoptimized link state routing protocol (OLSR)critical-section problem
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved