Deck 6: Storage and Other Io Topics

Full screen (f)
exit full mode
Question
What is the average time to read a 512 byte sector for Seagate ST31000340NS in Figure 6.5? What is the minimum time? Assume that the controller overhead is 0.2 ms, and the disk is idle so that there is no waiting time.
Use Space or
up arrow
down arrow
to flip the card.
Question
What is the bottleneck in the following system setup, the CPU, memory bus, or the disk set?
\bullet The user program continuously performs reads of 64KB blocks, and requires 2 million cycles to process each block.
\bullet The operating system requires 1 million cycles of overhead for each I/O operation.
\bullet The clock rate is 3GHz.
\bullet The maximum sustained transfer rate of the memory bus is 640MB/sec
\bullet The read/write bandwidth of the disk controller and the disk drives is 64MB/sec, disk average seek plus rotational latency is 9ms.
\bullet There are 20 disks attached to the bus each with its own controller. (Assume that each disk can be controlled independently and ignore disk conflicts.)
Question
In Figure 6.8 which fields are correlated with each other? Why do these correlations exist?
Question
In Figure 6.9, PCI-E connections are available from both the north bridge and the south bridge. What are the advantages and disadvantages to attaching devices to the PCI-E connections on the north and south bridges?
Question
Which of the following would be an acceptable transport medium for real-time transmission of human voice data? Which would be "overkill"?
· 56.5Kbps modem
· 100 Base-T Ethernet connection
· 802.11b wireless connection.
Question
Discuss why RAID 3 is not suited for transaction processing applications. What kind of applications is it suitable for and why?
Question
Imagine that you are proposing a new disk IO benchmark for transaction processing, what sort of experiments would you perform to show that your benchmark's results are meaningful. Imagine that you are reviewing a paper introducing a new disk IO benchmark for transaction processing. What sort of subtle flaws would you search for?
Question
Suppose we have two different I/O system A and B. A has data transfer rate: 5KB/s and has access delay: 5 sec. While B has data transfer rate: 3 KB/s and has access delay: 4 sec. Now we have a 3M I/O request, taking performance into consideration, which I/O system will you use? What about for a 3KB request?
Question
75 that hard disk block a+1 will be read soon afterwards. Given that an application has requested a read of a single disk block, the OS can expect the application to read subsequent blocks later. If the OS will pursue a strategy of reading n blocks at a time, analyze how the OS can choose this n in order to minimize the expected read time.
Question
7 before the first wear out if wear leveling working ideally?
Question
If a system contains 1,000 disk drives, and each of them has a 800,000 hour MTBF, how often a drive failure will occur in that disk system? Could you give some idea to improve that? And why will your idea work?
Question
A given computer system includes a hard disk with direct memory access (DMA).
(a) Suppose a user application needs to change a single byte within a disk block. Sketch, in order, all communications that must take place between the processor and the hard drive to complete this operation.
Question
Answer the following questions:
(a) What is the average time to read or write a 512-byte sector for a typical disk rotating at 7200 RPM? The advertised average seek time is 8ms, the transfer rate is 20MB/sec, and the controller overhead is 2ms. Assume that the disk is idle so that there is no waiting time.
(b) A program repeatedly performs a three-step process: It reads in a 4-KB block of data from disk, does some processing on that data, and then writes out the result as another 4-KB block elsewhere on the disk. Each block is contiguous and randomly located on a single track on the disk. The disk drive rotates at 7200RPM, has an average seek time of 8ms, and has a transfer rate of 20MB/sec. The controller overhead is 2ms. No other program is using the disk or processor, and there is no overlapping of disk operation with processing. The processing step takes 20 million clock cycles, and
the clock rate is 400MHz. What is the overall speed of the system in blocks processed per second assuming no other overhead?
Question
Mention two advantages and disadvantages for using a single bus as a shared communication link between memory, processor and I/O devices.
Question
Give a concise answer to each of the following questions. Limit your answers to 20-30 words.
(a) What is memory mapped I/O?
(b) Why is DMA an improvement over CPU programmed I/O?
(c) When would DMA transfer be a poor choice?
Question
Section 6.7 focuses on transactional processing as an example of a disk IO intensive application. Give another example of a disk IO intensive application compare and contrast the performance requirements and consider how different disk implementations (magnetic media, flash memory, or MEMS device) can be more or less appropriate for different applications.
Question
Disk Technology. Suppose we have a magnetic disk (resembling an IBM Microdrive) with the following parameters:
Disk Technology. Suppose we have a magnetic disk (resembling an IBM Microdrive) with the following parameters:   Answer the following questions. (Note: you may leave any answer as a fraction.) (a) What is the average time to read a single sector? (b) What is the average time to read 8 KB in 16 consecutive sectors in the same cylinder? (c) Now suppose we have an array of 4 of these disks. They are all synchronized such that the arms on all the disks are always on the same sector within the track. The data is striped across the 4 disks so that 4 logically consecutive sectors can be read in parallel. What is the average time to read 32 consecutive KB from the disk array?<div style=padding-top: 35px> Answer the following questions. (Note: you may leave any answer as a fraction.)
(a) What is the average time to read a single sector?
(b) What is the average time to read 8 KB in 16 consecutive sectors in the same cylinder?
(c) Now suppose we have an array of 4 of these disks. They are all synchronized such that the arms on all the disks are always on the same sector within the track. The data is striped across the 4 disks so that 4 logically consecutive sectors can be read in parallel. What is the average time to read 32 consecutive KB from the disk array?
Unlock Deck
Sign up to unlock the cards in this deck!
Unlock Deck
Unlock Deck
1/17
auto play flashcards
Play
simple tutorial
Full screen (f)
exit full mode
Deck 6: Storage and Other Io Topics
1
What is the average time to read a 512 byte sector for Seagate ST31000340NS in Figure 6.5? What is the minimum time? Assume that the controller overhead is 0.2 ms, and the disk is idle so that there is no waiting time.
Average =
Average seek time + Average rotational delay + Transfer time + controller overhead
= 8.5 + 0.5 rotations/7200RPM + 0.5KB/105MBps + 0.2
= 8.5+4.17+0.005+0.2
= 12.875ms
Minimum:
Minimum seek time + Minimum rotational delay + Transfer time + controller overhead
= 0.8 + 0 rotations + 0.5KB/105MBps + 0.2
= 0.8+0+0.005+0.2
= 1.005ms
2
What is the bottleneck in the following system setup, the CPU, memory bus, or the disk set?
\bullet The user program continuously performs reads of 64KB blocks, and requires 2 million cycles to process each block.
\bullet The operating system requires 1 million cycles of overhead for each I/O operation.
\bullet The clock rate is 3GHz.
\bullet The maximum sustained transfer rate of the memory bus is 640MB/sec
\bullet The read/write bandwidth of the disk controller and the disk drives is 64MB/sec, disk average seek plus rotational latency is 9ms.
\bullet There are 20 disks attached to the bus each with its own controller. (Assume that each disk can be controlled independently and ignore disk conflicts.)
Amount of time CPU takes to process each 64 KB block
= ((2 * 106)/ (3 * 109)) * 103
= 0.67 ms
Amount of time spent in memory transfer for 64 KB block
= ((64 * 210)/(640*220)) * 1000 = 0.097 ms
Amount of time spent in I/O transfer for a 64 KB block
= seek time + rotational delay + transfer time + controller overhead
= 9 + ((64*210)/(64*220))*103 + ((106 * 3)/109)*103
= 12.97 ms
The main bottleneck is I/O in the above system.
3
In Figure 6.8 which fields are correlated with each other? Why do these correlations exist?
Correlation 1: Each external bus supports multiple devices per channel whereas each internal device supports only a single device per channel. The reason for this correlation is that internal devices are rarely added by end users and can be only few in number due to space constraints. By contrast many external devices may be added although a computer can only have a few external ports. Outside of a computer there is enough room for additional hub devices.
Correlation 2: External devices have longer maximum bus length.
4
In Figure 6.9, PCI-E connections are available from both the north bridge and the south bridge. What are the advantages and disadvantages to attaching devices to the PCI-E connections on the north and south bridges?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
5
Which of the following would be an acceptable transport medium for real-time transmission of human voice data? Which would be "overkill"?
· 56.5Kbps modem
· 100 Base-T Ethernet connection
· 802.11b wireless connection.
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
6
Discuss why RAID 3 is not suited for transaction processing applications. What kind of applications is it suitable for and why?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
7
Imagine that you are proposing a new disk IO benchmark for transaction processing, what sort of experiments would you perform to show that your benchmark's results are meaningful. Imagine that you are reviewing a paper introducing a new disk IO benchmark for transaction processing. What sort of subtle flaws would you search for?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
8
Suppose we have two different I/O system A and B. A has data transfer rate: 5KB/s and has access delay: 5 sec. While B has data transfer rate: 3 KB/s and has access delay: 4 sec. Now we have a 3M I/O request, taking performance into consideration, which I/O system will you use? What about for a 3KB request?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
9
75 that hard disk block a+1 will be read soon afterwards. Given that an application has requested a read of a single disk block, the OS can expect the application to read subsequent blocks later. If the OS will pursue a strategy of reading n blocks at a time, analyze how the OS can choose this n in order to minimize the expected read time.
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
10
7 before the first wear out if wear leveling working ideally?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
11
If a system contains 1,000 disk drives, and each of them has a 800,000 hour MTBF, how often a drive failure will occur in that disk system? Could you give some idea to improve that? And why will your idea work?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
12
A given computer system includes a hard disk with direct memory access (DMA).
(a) Suppose a user application needs to change a single byte within a disk block. Sketch, in order, all communications that must take place between the processor and the hard drive to complete this operation.
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
13
Answer the following questions:
(a) What is the average time to read or write a 512-byte sector for a typical disk rotating at 7200 RPM? The advertised average seek time is 8ms, the transfer rate is 20MB/sec, and the controller overhead is 2ms. Assume that the disk is idle so that there is no waiting time.
(b) A program repeatedly performs a three-step process: It reads in a 4-KB block of data from disk, does some processing on that data, and then writes out the result as another 4-KB block elsewhere on the disk. Each block is contiguous and randomly located on a single track on the disk. The disk drive rotates at 7200RPM, has an average seek time of 8ms, and has a transfer rate of 20MB/sec. The controller overhead is 2ms. No other program is using the disk or processor, and there is no overlapping of disk operation with processing. The processing step takes 20 million clock cycles, and
the clock rate is 400MHz. What is the overall speed of the system in blocks processed per second assuming no other overhead?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
14
Mention two advantages and disadvantages for using a single bus as a shared communication link between memory, processor and I/O devices.
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
15
Give a concise answer to each of the following questions. Limit your answers to 20-30 words.
(a) What is memory mapped I/O?
(b) Why is DMA an improvement over CPU programmed I/O?
(c) When would DMA transfer be a poor choice?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
16
Section 6.7 focuses on transactional processing as an example of a disk IO intensive application. Give another example of a disk IO intensive application compare and contrast the performance requirements and consider how different disk implementations (magnetic media, flash memory, or MEMS device) can be more or less appropriate for different applications.
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
17
Disk Technology. Suppose we have a magnetic disk (resembling an IBM Microdrive) with the following parameters:
Disk Technology. Suppose we have a magnetic disk (resembling an IBM Microdrive) with the following parameters:   Answer the following questions. (Note: you may leave any answer as a fraction.) (a) What is the average time to read a single sector? (b) What is the average time to read 8 KB in 16 consecutive sectors in the same cylinder? (c) Now suppose we have an array of 4 of these disks. They are all synchronized such that the arms on all the disks are always on the same sector within the track. The data is striped across the 4 disks so that 4 logically consecutive sectors can be read in parallel. What is the average time to read 32 consecutive KB from the disk array? Answer the following questions. (Note: you may leave any answer as a fraction.)
(a) What is the average time to read a single sector?
(b) What is the average time to read 8 KB in 16 consecutive sectors in the same cylinder?
(c) Now suppose we have an array of 4 of these disks. They are all synchronized such that the arms on all the disks are always on the same sector within the track. The data is striped across the 4 disks so that 4 logically consecutive sectors can be read in parallel. What is the average time to read 32 consecutive KB from the disk array?
Unlock Deck
Unlock for access to all 17 flashcards in this deck.
Unlock Deck
k this deck
locked card icon
Unlock Deck
Unlock for access to all 17 flashcards in this deck.