Top Banner
Minimize Human Intervention Fault Tolerant Patterns. Architectural Patterns EVA Community Ji-won Kim([email protected]) 2013.8.24 : People make mistakes and are slow; imize downtime the system should take care of itself, t human intervention.
11

Fault tolerant 4_5

Jun 23, 2015

Download

Technology

Eva

Fault Tolerant Patterns, Essential Pattern, Principle.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Fault tolerant 4_5

MinimizeHumanIntervention

@ Fault Tolerant Patterns.Architectural Patterns

EVA Community Ji-won Kim([email protected]) 2013.8.24

Intent : People make mistakes and are slow; to minimize downtime the system should take care of itself, without human intervention.

Page 2: Fault tolerant 4_5

2

Story Line

Human is not a god

Prevent human from going wrong way

Hardware and Software is innocent

Communication

Debug?, Dehuman! On ( ? )

Page 3: Fault tolerant 4_5

Incomplete Human

50%~

25%

25%

The cause of failures

Human Hardware

Software

(From a study of the US telephone network[Kuh97])

Page 4: Fault tolerant 4_5

Human can do every-thing

Recognition/Report

Fix/Modification Analysis Destruction

Test/ PlanningOperation

Page 5: Fault tolerant 4_5

5

Human's Creation

Network

ComputingMachine

Math

Switch

Human is not a machine

AlwaysIdentical ProceduralSoftware&Hardware

Become Bored, inattentive with routine, monotonous tasks

Page 7: Fault tolerant 4_5

7

What’s the Prob-lem?

The risk of Procedural Errors

Incorrect system from imperfect Human’s

Requirements

Operator

Page 8: Fault tolerant 4_5

Whole System

Implementation &Operation

If it’s already planned and predictable, make it all automatic

8

Fault Observe

r

Monitoring System

Component1 Component2

Component3

Fault Observe

r

Fault Observe

r

RecoverySystem

Operator

Page 9: Fault tolerant 4_5

Benefit

Fast Response time for Error situationQuickly Recover ErrorsAvoid this

From Book, originally from©iStockphoto.com/Don Bayley

Page 10: Fault tolerant 4_5

Pattern Relation

Minimize Human In-tervention

Detection• Fault Ob-

server• Audible

Alarm

Recovery• Recovery

Blocks• Error Han-

dlers• Maximize

Human Par-ticipation

Management• Mainte-

nance In-terfaces

• IO Triage

Prevention& Correct• Reintegra-

tion• Revise Pro-

cedure

Helpful Patterns :An Input and Output Pattern Language [HS00]

Page 11: Fault tolerant 4_5

Risk?

Mis-operation Windows System Recovery