Framework Dependability - 3.


63 views
Uploaded on:
Category: People / Lifestyle
Description
manual repair with auto re-config. 6. Equipment based Fault Tolerance ... manual repair with auto re-config. 5. no repair conceivable. 1. 14. Equipment based Fault ...
Transcripts
Slide 1

Framework Reliability - 3 Processor Problems Power-on perspectives Run-time issues Hardware-based Fault Tolerance Fault-tolerant structures Matching structures to prerequisites

Slide 2

Processor Problems Techniques to answer these inquiries: Is it OK to utilize the processor framework by any stretch of the imagination? Is it safe to keep utilizing it as a part of its typical running mode? In the event that issues are recognized, would we be able to recuperate? We need quite far smooth recuperation from disappointments Power-on Aspects CPU Tests Verify processor accurately executes its scope of directions Ensure condition banners can be set and read effectively Check on-chip registers and so on can be composed and read accurately Verify numeric coprocessors You can compose your own code to play out these test seeing that appropriate to your application code

Slide 3

Processor Problems Power-on Aspects ROM Tests ROM substance could get to be undermined by EM beats/radiation Flash can be "incidentally" reconstructed Write tests to check for legitimacy on catalyst Run-time Issues Stack flood Inhibit hinders until current one is adjusted (not generally plausible) Service the interruptr before the following arrives (not generally conceivable) Monitor with programming Hardware-based observing (eg ARM)

Slide 4

Processor Problems Run-time Issues Corruption of basic variables Causes by electrical clamor, power supply changes, … Use check summing Make a (check summed duplicate), contrast and unique Make 3 duplicates, use greater part voting Instruction pointer defilement PC may indicate unprogrammed areas Fill with No-operation codes or it might indicate modified memory areas More hard to manage

Slide 5

Hardware-based Fault Tolerance operational necessities constant operation required non-persistent operation adequate diminished execution satisfactory must be returned rapidly to full administration full execution required stops & locks out securely come up short operational high-accessibility fall flat dynamic safeguard no repair conceivable robotized repair reasonable programmed re-config manual repair without re-config manual repair with auto re-config manual repair with manual recuperation Automatic reconfig with programmed recuperation

Slide 6

Hardware-based Fault Tolerance Fault-tolerant structures Passive excess frameworks Sensor 1 Processor 1 Processor 2 Majority voter Control actuator Processor 3 Sensor 2 Processor 4

Slide 7

Hardware-based Fault Tolerance Fault-tolerant structures Active repetitive frameworks Standby processor could run "chilly" or "hot", in the last case maybe completely synchronized with the dynamic Processor 1 (dynamic) Processor 3 (issue recognition) Sensors Processor 2 (standby) yield Bus switch

Slide 8

Hardware-based Fault Tolerance Fault-tolerant structures Hybrid repetitive frameworks Majority voter Processor 1 (dynamic) Processor 2 (dynamic) Sensor 1 Switching circuit Bus switch Processor 3 (dynamic) yield Processor 4 (save)

Slide 9

Hardware-based Fault Tolerance Fault-tolerant structures Very high accessibility frameworks Active excess structures DB servers and so on Very quick "failover" times Ready accessibility of circle information to standby frameworks Hot-swap frameworks Replacement of loads up while framework is on-line, without irritating operations CompactPCI standard: four levels of ability Basic hot-swap Full hotswap High-accessibility hot-swap (single processor) High-accessibility hot-swap (numerous processor)

Slide 10

Hardware-based Fault Tolerance Fault-tolerant structures CompactPCI - Basic hot-swap Devices, systems Host Processor single load up PC (SBC) Hot-swap I/O loads up PCI transport

Slide 11

Hardware-based Fault Tolerance Fault-tolerant structures CompactPCI - Full & high-accessibility (single processor) hot-swap Full: interface to framework programming permitting it to regulate swap of loads up HA: framework screens/controls loads up with stage administration programming Platform administration controller SBC Devices, systems Hot-swap I/O loads up PCI transport administration transport Host Processor SBC

Slide 12

Hardware-based Fault Tolerance Fault-tolerant structures CompactPCI - Full & high-accessibility (different processor) hot-swap Distributed figuring: load adjusting Re-portion of work from a fizzled processor Fault finding, repair oversaw by the host SBC Primary (host) SBC Devices, systems Dedicated I/O controller SBC 1 Devices, systems SBC 2 PCI transport SBC n

Slide 13

Hardware-based Fault Tolerance operational prerequisites ceaseless operation required non-nonstop operation worthy decreased execution satisfactory must be returned rapidly to full administration full execution required stops & locks out securely come up short operational high-accessibility come up short dynamic safeguard no repair conceivable 1 computerized repair admissible 2 programmed re-config 3 manual repair without re-config 4 manual repair with auto re-config 5 manual repair with manual recuperation 6 Automatic reconfig with programmed recuperation 7

Slide 14

Hardware-based Fault Tolerance Matching structures to Requirements

Recommended
View more...