You are designing a new distributed file system. Your file system lives in user space and files can grow to be very large. Files are composed of physically non-contiguous blocks. You are trying to decide how big to make your file blocks (This will become a fixed system parameter).

Your expected use cases include reading very large files, so you figure that if you make the transfer time of each block significantly larger than the seek time to the beginning of the block, you will be reading large files (made of multiple blocks) almost at the disk transfer rate.

If you want to make the disk seek time 1% of the block transfer time, approximately, how big should your blocks be? In other words, what will be your chosen block size? Also, when choosing the block size, take into account any other practical considerations (hint: think about whats a good size to round to).

Explain your calculation.

You may use any number that we covered in class, or any other number you research, but make sure to explicitly specify what numbers you are using, and any assumption you make.

Avoid the following pitfall: when research disk transfer rate, take into account whether the stated number is in bits per second (bps) or bytes per seconds, and make the proper adjustments.