My solution to problem 1 ^^^

1. Consider the setting of a machine with a single disk, with the following properties: a. The throughput is 256MB/sec. b. There is a file of size 500GB How long does it take to read the entire file into the main memory? 500 GB = 512,000 MB 512,000 MB = 2,000 sec 256- 2,000 sec 33 mins MB sec Now, consider distributed setting, where the cluster has the following properties: a. Each node has the same throughput (256MB/sec.). b. There is a file of size 500GB which is split into blocks, and each block is 128MB. c. There are a total of 5000 nodes. What would be the best-case scenario of distributing the blocks among the nodes, for the purpose of optimizing the throughput (in terms of making sure that the entire file is read (i.e., each portion is read from disk -> main memory)? What is the speed-up with respect to Problem 1? 1. Consider the setting of a machine with a single disk, with the following properties: a. The throughput is 256MB/sec. b. There is a file of size 500GB How long does it take to read the entire file into the main memory? 500 GB = 512,000 MB 512,000 MB = 2,000 sec 256- 2,000 sec 33 mins MB sec Now, consider distributed setting, where the cluster has the following properties: a. Each node has the same throughput (256MB/sec.). b. There is a file of size 500GB which is split into blocks, and each block is 128MB. c. There are a total of 5000 nodes. What would be the best-case scenario of distributing the blocks among the nodes, for the purpose of optimizing the throughput (in terms of making sure that the entire file is read (i.e., each portion is read from disk -> main memory)? What is the speed-up with respect to Problem 1