(1). Assume we have a process "ready" list that has 8 levels of priority. Assume also that the dispatcher selects highest priority processes first, but randomly selects processes from any single priority level (i.e. if there are 2 processes at the highest level, either can be chosen). If there are 5 processes only in the list, all at level 4, how many possible dispatcher schedules are there if no new processes are added? What if two of the processes are at level 2 and the rest at level 4?
(2). Of the three disk allocation methods used for file system implementation, which would be the easiest method for the coding of the "lseek" command? Hardest? Explain.
(3). The "public key" encryption system at its best, uses 100+ digit (not bit) prime numbers as keys. The online credit card info submission only uses a 15 digit key which is much more easily breakable yet we still use it. Why?
(4). Assume that we have a disk with 200 tracks, numbered 0-199, which is serving a request at track 123 and has just finished a request at track 115. The queue of requests is kept in the following FIFO order:
147, 95, 174, 94, 128, 150, 102, 171, 111, 86, 120, 162, 70
Show the order of service (assuming no more requests) for SSTF, SCAN, and C-SCAN disk-scheduling algorithms.
(5). When employing resourcs preemption in a deadlock recovery situation, there is a counter for each process which is incremented when that process is selected as a "victim" by the mid-level scheduler. What is the purpose of this counter?
(6). Consider datagram sockets and stream sockets. Which would be easiest to implement? Why?
(7). Consider the following page reference string:
1, 2, 3, 4, 7, 1, 5, 6, 2, 1, 4, 3, 2, 1, 2, 3, 6, 7,
(yes, you've seen it before) How many page faults would occur for both the LRU and FIFO replacement algorithm assuming a FIVE frame working set?
(8). What is the relationship between inodes and vnodes?
(9). In a multi-level indexed file system that has 1K blocks what is the maximum file size if the i-nodes contain 8 direct entries, one single and one double indirect entry each?
(10). In a distributed system that may use either a centralized or distributed approach to mutual exclusion, what impact is experienced when we have station crashes? If station crashes are common, which approach would you use? Why?