CS-3310 - 04 I/O and Disks - Disks

data density of typical hard disk (video)

31B bits per cm^2

head above rotating disk is like (video)

747 with 100,000 passengers flying at 60 mph at 0.01 in above ground

how many data accesses per second on a typical hard disk? (video)

data access up to 60x per second

What is the basic structure of a hard disk?

One or more spinning platters coated with magnetic material, with read/write heads on each surface.

What is a platter?

A circular disk surface where data is stored magnetically.

What is a track?

A concentric ring of sectors on a platter.

What is a sector?

The smallest addressable block of data on a disk, typically 512 bytes.

What is a cylinder?

The set of tracks across all platters that align vertically.

What are the main components of disk access time?

Seek time, rotational delay, and transfer time.

What is seek time?

The time to move the disk arm to the correct track.

What is rotational delay?

The time for the desired sector to rotate under the read/write head.

What is transfer time?

The time to actually read or write the data once positioned.

How is total I/O time calculated?

I/O time = Seek time + Rotational delay + Transfer time.

Why is disk scheduling needed?

To reduce average access time and improve throughput by reordering I/O requests.

What does the SSTF (Shortest Seek Time First) algorithm do?

Chooses the pending request closest to the current head position.

What is the main disadvantage of SSTF?

It can cause starvation for distant requests.

How does the SCAN (Elevator) algorithm work?

It moves the disk head back and forth, servicing all requests in one direction before reversing.

What is C-SCAN?

A variation of SCAN that only moves in one direction, wrapping around to the beginning for fairness.

Why does C-SCAN provide more uniform response times than SCAN?

Because it treats all requests equally regardless of position on the disk.

What is SPTF (Shortest Positioning Time First)?

A scheduling method that minimizes total access time by considering both seek and rotational delay.

Who typically performs SPTF scheduling?

The disk controller firmware, not the OS.

What is RAID?

Redundant Array of Inexpensive Disks - a method of combining multiple disks for performance and reliability.

What are the main goals of RAID?

Improved performance, increased capacity, and better reliability.

Describe RAID 0.

Striping data across multiple disks for performance; no redundancy.

Describe RAID 1.

Mirroring data on multiple disks for reliability.

Describe RAID 4 and RAID 5.

Both use parity for fault tolerance; RAID 5 rotates parity among disks to balance load.

What is double parity RAID (RAID 6)?

A scheme that can tolerate two disk failures.

What is low-level disk formatting?

The process of dividing each platter into tracks and sectors with preambles, ECC, and identifiers.

What is high-level formatting?

Creating a file system structure like root directory, free list, and superblock.

What are MBR and GPT?

Partition table schemes - MBR for small disks (<2TB), GPT for larger disks.

How do disks handle bad sectors?

By remapping them to spare sectors stored on the drive.

What is the purpose of ECC in disks?

Error Correction Code helps recover data from partially damaged sectors.