Free Response Question 1: In preparation for the Silurian in…
Free Response Question 1: In preparation for the Silurian invasion, the Lizard Legion maintains a database tracking all persons of interest on Earth. This database is hosted on the Basilisk Hypercomputer, a multiuser system that the Silurian overseers use for their daily tasks. Recently, the Legion has redoubled its efforts to maintain and update the database. Unfortunately, this has somehow resulted in a catastrophic reduction in work efficiency for users on the Basilisk Hypercomputer. The Silurian High Council, furious at the disruption and chaos, tasks you with solving the problem. In your initial research, you note that the lack of work efficiency is due to Silurian overseers spending hours twiddling their digits in front of frozen workstations; the overseers you speak to claim they are attempting to access the database. You decide to look into the database program and learn that every time a user on the Basilisk Hypercomputer accesses the database, that user acquires exclusive access. Any other users requesting access before the first user is finished acquire a spinlock/perform busy waiting until it is their turn. A) What exactly is the issue causing the frozen workstations? Explain your reasoning. B) Explain how you might modify the database program to prevent this issue. C) What are the downsides or consequences of your solution? Free Response Question 2: The Lizard Legion just finished implementing a high-speed network system. Unfortunately, their system runs much slower than expected! The intended use case for the system is to process incoming packets. Users of the system will produce packets to be sent to a holding chamber, where multiple servers accept and process these packets concurrently. Currently, when a user sends a packet to the holding chamber, all other users must wait. On top of that, it seems like multiple servers are processing the same packet at the same time, even though after a packet is processed, the server removes it from the holding chamber. A) What type of synchronization problem is this? B) Why are multiple servers processing the same packet at the same time, even though the server removes the packet from the holding chamber once processed? How do we fix this? C) How can we optimize the system so that multiple users can send packets to the holding chamber at the same time? What concerns should we keep track of regarding safety and concurrency?