CMPUT429/CMPE382 Winter 2001

3/22/02


Click here to start


Table of Contents

CMPUT429/CMPE382 Winter 2001

Motivation: Who Cares About I/O?

I/O Systems

Outline

Disk Device Terminology

Photo of Disk Head, Arm, Actuator

Disk Device Performance

Disk Device Performance

Data Rate: Inner vs. Outer Tracks

Devices: Magnetic Disks

Disk Performance Model /Trends

State of the Art: Barracuda 180

Disk Performance Example (will fix later)

Areal Density

Areal Density

MBits per square inch: DRAM as % of Disk over time

Historical Perspective

Disk History

Disk History

1 inch disk drive!

Disk Characteristics in 2000

Disk Characteristics in 2000

Disk Characteristics in 2000

Disk Characteristics in 2000

Fallacy: Use Data Sheet “Average Seek” Time

Fallacy: Use Data Sheet Transfer Rate

Disk Performance Example

Future Disk Size and Performance

Tape vs. Disk

Current Drawbacks to Tape

Automated Cartridge System: StorageTek Powderhorn 9310

Library vs. Storage

Whither tape?

Use Arrays of Small Disks?

Advantages of Small Formfactor Disk Drives

Replace Small Number of Large Disks with Large Number of Small Disks! (1988 Disks)

Array Reliability

Redundant Arrays of (Inexpensive) Disks

Redundant Arrays of Inexpensive Disks RAID 1: Disk Mirroring/Shadowing

Redundant Array of Inexpensive Disks RAID 3: Parity Disk

RAID 3

Inspiration for RAID 4

Problems of Disk Arrays: Small Writes

System Availability: Orthogonal RAIDs

System-Level Availability

Berkeley History: RAID-I

Summary: RAID Techniques: Goal was performance, popularity due to reliability of storage

Summary Storage

Reliability Definitions

IFIP Standard terminology

Fault v. (Latent) Error v. Failure

Fault v. (Latent) Error v. Failure

Fault v. (Latent) Error v. Failure

Fault v. (Latent) Error v. Failure

Fault Tolerance vs Disaster Tolerance

Defining reliability and availability quantitatively

Fail-Fast is Good, Repair is Needed

Dependability: The 3 ITIES

Reliability Example

What's wrong with MTTF?

What's wrong with MTTF?

Dependability Big Idea: No Single Point of Failure

HW Failures in Real Systems: Tertiary Disks

When To Repair?

Sources of Failures

Case Study - Japan "Survey on Computer Security", Japan Info Dev Corp., March 1986. (trans: Eiichi Watanabe).

Case Studies - Tandem Trends Reported MTTF by Component

Is Maintenance the Key?

OK: So Far

Does Hardware Fail Fast? 4 of 384 Disks that failed in Tertiary Disk

High Availability System Classes Goal: Build Class 6 Systems

How Realistic is "5 Nines"?

Summary: Dependability

Summary: Dependability

Introduction to Queueing Theory

A Little Queuing Theory: Notation

A Little Queuing Theory

A Little Queuing Theory

A Little Queuing Theory: Variable Service Time

A Little Queuing Theory: Average Wait Time

A Little Queuing Theory: M/G/1 and M/M/1

A Little Queuing Theory: An Example

A Little Queuing Theory: Another Example

A Little Queuing Theory: Another Example

A Little Queuing Theory: Yet Another Example

Pitfall of Not using Queuing Theory

Author: Randy H. Katz

Email: amaral@cs.ualberta.ca

Home Page: www.cs.ualberta.ca

Download presentation source