-
Hints and Answers Your friend has just bought a new personal computer. She tells you that her new system runs at 1GHz, which makes it over three times faster than her old 300MHz system. What would you tell her?
-
Suppose the daytime processing load consists of 60% CPU activity and 40% disk activity. Your customers are complaining that the system is slow. After doing some research, you learn that you can upgrade your disks for $8,000 to make them 2.5 times as fast as they are currently. You have also learned that you can upgrade your CPU to make it 1.4 times as fast for $5,000.
-
Which would you choose to yield the best performance improvement for the least amount of money?
-
Which option would you choose if you don't care about the money, but want a faster system?
-
What is the break-even point for the upgrades? That is, what price would be charged for both upgrades to make their cost and performance improvement equal?
-
Hints and Answers How would you answer Question 2 if the system activity consists of 55% processor time and 45% disk activity?
-
Name the four types of I/O architectures. Where are each of these typically used and why are they used there?
-
A CPU with interrupt-driven I/O is busy servicing a disk request. While the CPU is midway through the disk-service routine, another I/O interrupt occurs.
-
Hints and Answers What happens next?
-
Hints and Answers Is it a problem?
-
Hints and Answers If not, why not? If so, what can be done about it?
-
Why are I/O buses provided with clock signals?
-
If an address bus needs to be able to address eight devices, how many conductors will be required? What if each of those devices also needs to be able to talk back to the I/O control device?
-
We pointed out that I/O buses do not need separate address lines. Construct a timing diagram similar to Figure 7.7 that describes the handshake between an I/O controller and a disk controller for a write operation. (Hint: You will need to add a control signal.)
-
Asterisk Mark If each interval shown in Figure 7.7 is 50 nanoseconds, how long would it take to transfer 10 bytes of data? Devise a bus protocol, using as many control lines as you need, that would reduce the time required for this transfer to take place. What happens if the address lines are eliminated and the data bus is used for addressing instead? (Hint: An additional control line may be needed.)
-
Define the terms seek time, rotational delay, and transfer time. Explain their relationship.
-
Hints and Answers Why do you think the term random access device is something of a misnomer for disk drives?
-
Why do differing systems place disk directories in different track locations on the disk? What are the advantages of using each location that you cited?
-
Hints and Answers Verify the average latency rate cited in the disk specification of Figure 7.11. Why is the calculation divided by 2?
-
By inspection of the disk specification in Figure 7.11, what can you say about whether the disk drive uses zoned-bit recording?
-
The disk specification in Figure 7.11 gives a data transfer rate of 6.0MB per second when reading from the disk, and 11.1MB per second when writing to the disk. Why are these numbers different?
-
Do you trust disk drive MTTF figures? Explain.
-
Suppose a disk drive has the following characteristics:
-
Hints and Answers What is the capacity of the drive?
-
Hints and Answers What is the access time?
-
Suppose a disk drive has the following characteristics:
-
What is the capacity of the drive?
-
What is the access time?
-
Is this disk faster than the one described in Question 17? Explain.
-
What are the advantages and disadvantages of having a small number of sectors per disk cluster?
-
Asterisk Mark Suggest some ways in which the performance of a 1.44MB floppy disk could be improved.
-
What is the maximum number of root directory entries on a 1.44MB floppy? Why?
-
How does the organization of an optical disk differ from the organization of a magnetic disk?
-
Discuss the difference between how DLT and DAT record data. Why would you say that one is better than the other?
-
How would the error-correction requirements of an optical document storage system differ from the error-correction requirements of the same information stored in textual form? What are the advantages offered by having different levels of error correction for optical storage devices?
-
You have a need to archive a large amount of data. You are trying to decide whether to use tape or optical storage methods. What are the characteristics of this data and how it is used that will influence your decision?
-
Asterisk Mark A particular high-performance computer system has been functioning as an e-business server on the Web. This system supports $10,000 per hour in gross business volume. It has been estimated that the net profit per hour is $1,200. In other words, if the system goes down, the company will lose $1,200 every hour until repairs are made. Furthermore, any data on the damaged disk would be lost. Some of this data could be retrieved from the previous night's backups, but the rest would be gone forever. Conceivably, a poorly timed disk crash could cost your company hundreds of thousands of dollars in immediate revenue loss, and untold thousands in permanent business loss. The fact that this system is not using any type of RAID is disturbing to you.
Although your chief concern is data integrity and system availability, others in your group are obsessed with system performance. They feel that more revenue would be lost in the long run if the system slowed down after RAID is installed. They have stated specifically that a system with RAID performing at half the speed of the current system would result in gross revenue dollars per hour declining to $5,000 per hour.
In total, 80% of the system e-business activity involves a database transaction. The database transactions consist of 60% reads and 40% writes. On average, disk access time is 20ms.
The disks on this system are nearly full and are nearing the end of their expected life, so new ones must be ordered soon. You feel that this is a good time to try to install RAID, even though you'll need to buy extra disks. The disks that are suitable for your system cost $2,000 for each 10 gigabyte spindle. The average access time of these new disks is 15ms with an MTTF of 20,000 hours and an MTTR of 4 hours. You have projected that you will need 60 gigabytes of storage to accommodate the existing data as well as the expected data growth over the next 5 years. (All of the disks will be replaced.)
-
Are the people who are against adding RAID to the system correct in their assertion that 50% slower disks will result in revenues declining to $5,000 per hour? Justify your answer.
-
What would be the average disk access time on your system if you decide to use RAID-1?
-
What would be the average disk access time on your system using a RAID-5 array with two sets of four disks if 25% of the database transactions must wait behind one transaction for the disk to become free?
-
Which configuration has a better cost-justification, RAID-1 or RAID-5? Explain your answer.
-
-
Which of the RAID systems described in this chapter cannot tolerate a single disk failure?
-
Which can tolerate more than one simultaneous disk failure?
-
Compute the compression factors for each of the JPEG images in Figure 7.30.
-
Create a Huffman tree and assign Huffman codes for the "Star Bright" rhyme used in Section 7.8.2. Use <ws> for whitespace instead of underscores.
-
Complete the LZ77 data compression illustrated in Section 7.8.2.
-
JPEG is a poor choice for compressing line drawings, such as the one shown in Figure 7.30. Why do you think this is the case? What other compression methods would you suggest? Give justification for your choice(s).
-
-
Name an advantage of Huffman coding over LZ77.
-
Name an advantage of LZ77 over Huffman coding.
-
Which is better?
-
State one feature of PNG that you could use to convince someone that PNG is a better algorithm than GIF.
