Q5: Consider a distributed system with a client/server architecture where "mobile clients" on unreliable networks submit requests to a distributed multi-node batch job server. Job requests run on individual server nodes as a single thread, long-running, sequential jobs. Due to the algorithmic nature, these jobs cannot be parallelized and must run continuously on a dedicated CPU core of a server node for a long time (10 minutes). After submitting requests, the clients remain connected to the server while waiting for their job complete. After analyzing system transaction logs, it is determined that while the server nodes are not overloaded, and all computing jobs successfully run to completion (there is plenty of computing capacity to process client requests), ONLY 57% OF JOB RESULTS ARE SUCCESSFULLY PROVIDED TO THE CLIENTS. Continue using mobile clients without modifying the network, suggest how the design of the system could be improved to increase the % of job result successfully provided to clients.