Disk

• Stack of Magnetic Platters
  • Rotate together on a central spindle @3600-15000 RPM
  • Drive drifts slowly over time
    • Cannot predict position after 100+ rotations
• Disk arm assembly
  • Arms rotate around pivot
    • All move together
  • Arms contain disk heads
    • one for each recording surface
    • Heads read data and write data to platters

Disk positioning system

Sectors

Cost model for disk I⧸O

Disk Interface

Disk Performance

Scheduling for Disks

Approach 1: First come first served

Advantages

• Easy to implement
• Good fairness

Disadvantages

• Cannot exploit locality
• Higher latency → Lower throughput

Approach 2: Shortest Positioning time first

Always pick request with shortest seek time

Advantages

• Exploits locality
• Higher throughput

Disadvantages

• Starvation
• Don’t always know what request will be fastest

Approach 3: Elevator Scheduling (SCAN)

Like SPTF but next seek must be in same direction

• Switch directions only if no further requests

Advantages

• Locality
• Bounded waiting

Disadvantages

• Cylinders in middle of disk get better service

VSCAN

Creates a continuum between SPTF and SCAN

• Like SPTF but introduce effective positioning time
  If the request is same direction → $T_{eff}=T_{pos}$
  Otherwise → $T_{eff}=T_{pos}+r\ast T_{max}$
• $T_{max}$ is maximum seek time
• Penalty for changing directions
  r is between 0 and 1
• r=0 is SPTF
  • Just choose closest
• r=1 you get SCAN
  • Never change
• r=0.2 works well