1. Suppose now your two RAIDs are a RAID 5 and a RAID 6. Each of these RAIDs has a read speed of 48 MB/sec. In the event of a disk failure in either of these RAIDs, read performance is degraded by 12% as data from the failed disk must now be re-constructed from the RAIDs redundant data. For each of the cases below, assume the 875 Kbytes of data to be read is in physically sequential sectors on the disk, so that no additional seeks or delays occur once reading begins.

(a) If both RAIDs perform normally (i.e., all drives working normally), how long does it take for the RAID 5 to read this data? for the RAID 6?

(b) Suppose one disk has failed in each of the arrays. How long does it take for the RAID 5 to read this data? for the RAID 6?

(c) Suppose a total of two disks have failed in each of the arrays. How long does it take for the RAID 5 to read this data? for the RAID 6?

2. Three years ago your company bought a new RAID 5 array. All the disks in the RAID were brand new, all manufactured within a month of each other. Today one disk has failed in this RAID. How many disks do you replace? Why?

3. Explain why its difficult to support direct access to files with variable length records. Suggest a method for handling this type of file if direct access is required.