Reliability of Computer Systems

Pregunta	Respuesta
The five things that can go wrong with a computer system	hardware failure software failure human error malicious damage natural disasters
Some of the consequences of non-life-critical systems failing	Loss of revenue Dissatisfied customers Damage to reputation Fines for breaking the law
Examples of critical systems	Utilities Supplies Banking Emergency services Telephone network Food supply Air traffic control Nuclear safety
Some consequences of critical systems failure	Risks lives of customers Put business to a halt
Malicious intent with a critical system	Cyber-terrorism Cyber terrorists are hard to catch and prosecute when operating in another country
Types of backing-up data	Full backup: makes a copy of everything Incremental backup: only makes a copy of new or edited files
Backup Procedure	A backup schedule should be in place and there should be someone in charge of it. Backup sites should be off of the main companies site in a locked, fireproof place
Disaster Recovery Definition	Getting a system back to normal after a large-scale failure.
Disaster Recovery Procedure	Most organizations have a disaster recovery plan. After a system failure, data must be restored from a backup.
Redundancy Definition	Have duplicate resources or extra resources that can support the main system. A backup system for if the main one goes down
Three main types of redundancy	Hardware like reserve storage or servers Software like a backup OS Data (already covered this)
Parallel Setup	Have two identical hardware setups that allow the company to switch between one and other, reducing disruption to their normal operations
Fault-tolerant Design	If parts of the system fail the system is designed to be able to continue running or 'fail gracefully' if it needs to completely stop
Failing Gracefully	The system will take steps to avoid data loss when things begin to go wrong. This could be writing an emergency copy of data before shutting down unexpectedly. This emergency copy is called a 'dump file'
Defensive Programming	Adding checks to see if the program is running properly. This can warn a user of a fault to prevent unexpected crashing
Measuring Reliability	Percentage of time working Number of hours operating without issues Downtime Mean time between failures Mean time to fail Image: Large (image/png)
Alpha Testing	Unit testing - checks a program's components Validation Testing - checks the type of data that can be used Volume testing - checks the amount off data a system can handle
Beta Testing	Users warn the programmers of any problem and the programs can be fixed and updated weekly or monthly for example
End-users	Users can find unexpected problems with a system. Updates are then released. Some companies reward people for finding problems e.g. Facebook

Siguiente

Descripción

Resumen del Recurso

Similar

	Creado por Josh Watkins hace más de 7 años