Problem 2 You have rigged a public file compression server. The server operates as follows. When a user uploads a file, is is stored locally and compressed. After compression a file deflates to 50% its original size. After compression is done, a download for the compressed file is initiated. After the download is completed, both the original file and the compressed file are deleted from the server. For doing this, you are using two machines. One of them is a low-power one that has a SSD drive with only 128 GB of space to store files The second one is an older machine that you bought second-hand from a sketchy company called StatWorks It has a much bigger drive (quite impossible to fill up) and a faster CPU, but it consumes 10x the power so you are trying to minimize its usage After some banchmarking, you have evaluated that the average file uploaded by users is 16 GB. And that on the low-power machine, it takes about 3 minutes from upload to completion of download when a single request is being handled (i.e. when there is no queuing). When the SSD is filled-up with files that await processing, you redirect users to the faster server, which takes about 1 minutes from upload to compressed file download completion again with no queuing. Moreover you have evaluated that every 8 minutes you receive about 2 file compression requests. Answer the following a) What is the likelihood that a user will be served by the high-power server? b) What is the average amount of space occupied in the high-power server? c) If the low-power server consumes 1 W when active and 0 W when idle, what is the average consumption of the entire system (including the high-power server)? d) How much space on average is being taken by user files on the SSD of the low-power server? e) What time estimate you provide to your users (regardless of the current system state, and without knowledge of the files they want to upload) about how long it will take before their compressed file download completes? Problem 2 You have rigged a public file compression server. The server operates as follows. When a user uploads a file, is is stored locally and compressed. After compression a file deflates to 50% its original size. After compression is done, a download for the compressed file is initiated. After the download is completed, both the original file and the compressed file are deleted from the server. For doing this, you are using two machines. One of them is a low-power one that has a SSD drive with only 128 GB of space to store files The second one is an older machine that you bought second-hand from a sketchy company called StatWorks It has a much bigger drive (quite impossible to fill up) and a faster CPU, but it consumes 10x the power so you are trying to minimize its usage After some banchmarking, you have evaluated that the average file uploaded by users is 16 GB. And that on the low-power machine, it takes about 3 minutes from upload to completion of download when a single request is being handled (i.e. when there is no queuing). When the SSD is filled-up with files that await processing, you redirect users to the faster server, which takes about 1 minutes from upload to compressed file download completion again with no queuing. Moreover you have evaluated that every 8 minutes you receive about 2 file compression requests. Answer the following a) What is the likelihood that a user will be served by the high-power server? b) What is the average amount of space occupied in the high-power server? c) If the low-power server consumes 1 W when active and 0 W when idle, what is the average consumption of the entire system (including the high-power server)? d) How much space on average is being taken by user files on the SSD of the low-power server? e) What time estimate you provide to your users (regardless of the current system state, and without knowledge of the files they want to upload) about how long it will take before their compressed file download completes
