What performance-related consequence is guaranteed to happen to average, assuming the remaining cluster nodes are fully operational?

An organization uses a four (4) node customer hosted Mule runtime cluster to host one (1) stateless api implementation. The API is accessed over HTTPS through a load balancer that uses round-robin for load distribution. Each node in the cluster has been sized to be able to accept four (4) times the current number of requests.

Two (2) nodes in the cluster experience a power outage and are no longer available. The load balancer directs the outage and blocks the two unavailable the nodes from receiving further HTTP requests.

What performance-related consequence is guaranteed to happen to average, assuming the remaining cluster nodes are fully operational?
A . 100% increase in the average response time of the API
B . 50% reduction in the throughput of the API
C . 100% increase in the number of requests received by each remaining node
D . 50% increase in the JVM heap memory consumed by each remaining node

Answer: A

Explanation:

* "100% increase in the throughput of the API" might look correct, as the number of requests processed per second might increase, but is it guaranteed to increase by 100%? Using 4 nodes will definitely increase throughput of system. But it is cant be precisely said if there would be 100% increase in throughput as it depends on many other factors. Also it is nowhere mentioned in the description that all nodes have same CPU/memory assigned. The question is about the guaranteed behavior

* Increasing number of nodes will have no impact on response time as we are scaling

application horizontally and not vertically. Similarly there is no change in JVM heap memory usage.

* So Correct answer is 50% reduction in the number of requests being received by each node This is because of the two reasons.

1) API is mentioned as stateless

2) Load Balancer is used

Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments