19. Disks

Question 1

What is stable storage?

Answer 1

Storage that does not lose its contents when the computer is turned off

Question 2

What are the two categories of stable storage?

Answer 2

HDD (hard disk drive), spinning disk, or hard drive

SSD (solid state drive) or flash drive

Question 3

Describe the hard drive category of stable storage.

Answer 3

A stable storage device constructed of rotating magnetic platters

Question 4

Describe the solid state drive category of stable storage.

Answer 4

A stable storage device constructed of non-moving, non-volatile memory

Question 5

What are the key differences between HDDs and SSDs?

Answer 5

HDDs are bigger, slower, and cheaper than SSDs.

Ex: HHD is $0.07 per GB, SSD is $0.15 per GB

Question 6

Why study spinning disks?

Answer 6

HDDs are still around and will be for awhile. Flash storage is definitely the future though.

Question 7

What are the three parts of the hardware disk?

Answer 7

Platter, Spindle, and Head

Question 8

What is the platter in a HDD?

Answer 8

A circular flat disk on which magnetic data is stored.

Constructed of a rigid non-magnetic material coated with a very thin layer of magnetic material.

Can have data written on both sides.

Question 9

What is the spindle in a HDD?

Answer 9

The drive shaft on which multiple platters are mounted and spun between 4200-15,000 RPM

Question 10

What is the head in a HDD?

Answer 10

The actuator that reads and writes data onto the magnetic surface of the platters while rotating at tens of nanometers over the platter surface

Question 11

What are the three components of disk location?

Answer 11

Track, Sector, and Cylinder

Question 12

What is the track?

Answer 12

Think of a lane on a race track running around the platter.

Question 13

What is a sector?

Answer 13

It resembles a slice of pie cut out of a single platter.

Question 14

What is a cylinder?

Answer 14

Imagine the intersection between a cylinder and the set of platters. It’s composed of a set of vertically-aligned tracks on all platters.

Question 15

How are disks different from other hardware components?

Answer 15

Disks move, disks are slow, and disks are devices, meaning they are less tightly coupled to the abstractions built on top of them.

Question 16

Describe the difference between the electronics time scale and the mechanics time scale.

Answer 16

The time necessary for electrons to flow from one part of the computer or chip to another is VERY fast.

The time necessary for a physical object to move gently and precisely from one point to another point is VERY slow comparatively.

Question 17

What do we mean when we say that disks can fail by parts?

Answer 17

Many disks ship with sectors that are already broken. The operating system - or disk software - detect and ignore these sectors.

Sectors may also fail over time, potentially resulting in data loss.

Question 18

What do we mean when we say that disks can fail catastrophically?

Answer 18

A head crash occurs when a jolt or violent event sends the disk heads crashing into the magnetic surface, scraping off material and destroying the platter.

When this happens, you have about 20 seconds to say goodbye.

Question 19

On the surface, what is RAID?

Answer 19

A clever way to use multiple disks to create a more reliable and better-performing device that looks like a single disk.

Question 20

How can the OS get directly involved to hide disk latency?

Answer 20

OS can use the past to predict the future, use a cache, or procrastination.

This is in contrast with how memory latencies hare handled. These are usually handled transparently by the processor through out-of-order execution.

Question 21

What are the 5 time-consuming steps of a read/write to disk?

Answer 21

1. Issue the command: OS has to tell device what to do and device selects which head to use
2. Seek time: Drive has to move the heads to appropriate track
3. Settle time: Heads have to stabilize on the (very narrow) track
4. Rotation time: Platters have to rotate to the position where the data is located
5. Transfer time: Data has to be read and transmitted back across the interconnect into system memory

Question 22

What improvements can we make on the harddrive to improve disk latencies?

Answer 22

Improve interconnect speeds (SATA-6?)

Improve seek times (Not improving rapidly. This is the moving physical object part).

Improve rotation speeds (This can vary between devices, but isn’t typically the major source of latency anyway. Physical limitations come into play here as well as to how fast the disk can spin.)

Question 23

Give a broad overview of the I/O “crisis” that arose across the 90s and 00s.

Answer 23

Hard drive densities and capacities soared, which encouraged users to save more stuff, thus increasing I/O demand.

At the same time, seek times limited the ability of disks to keep up with the added demand.

Capacity increased by three orders of magnitude, while speed only increased by two orders of magnitude.

Question 24

How is the file (disk) abstraction fundamentally different than the thread (CPU) and address space (memory) abstractions?

Answer 24

The file abstraction is almost entirely implemented in software.

Question 25

What is true about the nature of disk hardware that motivated the file abstraction?

Answer 25

The low-level disk interface is messy and very limited. It requires reading and writing entire 512-byte blocks without any notion of files or directories.

The file abstraction, through software, creates the ability to name, reference, and organize by directory data on the disk.

Question 26

What are two limitations of flash memory?

Answer 26

We have to erase entire large chunks before we can rewrite to that space.

Flash wears out faster than magnetic drives and can wear unevenly if certain areas are used more than others.