Load Balancers

Correct Answer: Adjust the load balancing algorithm weight for the underutilized server.

The best solution to the described problem is to adjust the load balancing algorithm weight for the underutilized server. This makes the load balancer direct more requests to this server, ensuring better utilization of resources and therefore improving application performance. The other answers are not as effective: applying aggressive timeout settings doesn't guarantee better utilization and might instead lead to lost or prematurely terminated requests. Disabling caching on the server doesn't directly address load balancing issues and may slow response times, and changing the load balancer's forwarding rules could affect the entire system negatively if not done carefully, without necessarily improving the situation with the underutilized server.

Correct Answer: Round Robin

The Round Robin algorithm should be used to distribute web traffic evenly across the three servers. This method dispenses requests to servers sequentially and cyclically, ensuring equal distribution. The 'Least Connections' algorithm is inappropriate if the three servers have different processing capabilities because it would forward the request to the server with the least connections regardless of its processing power. 'Fastest Response Time' would not ensure equal distribution as it would preferentially send requests to the fastest server, overloading it. The 'Random' algorithm would not guarantee even distribution over time because the traffic would be forwarded at random.

Correct Answer: Add the missing subnet to the load balancer configuration.

The problem mentioned in the question indicates that the users from a specific subnet are unable to access the application after the deployment of a load balancer. This suggests a possible misconfiguration in the subnet settings of the load balancer. Therefore, the most appropriate solution to fix the issue is to add the missing subnet to the load balancer configuration. This will ensure that the load balancer recognises the network of the affected users and correctly routes their data traffic to the application servers. Now, rebooting all servers connected to the load balancer is not necessary since the issue is likely due to misconfiguration rather than a server failure. Similarly, decreasing the load balance algorithm weight for the affected subnet won't address the current issue because the problem is not related to load balancing, but rather to network accessibility. Lastly, replacing application servers with more powerful hardware does not solve the current network accessibility issue and would be an unnecessary expense for the company.