Cloud - kamialie/knowledge_corner GitHub Wiki
Overview
Benefits:
- cost savings
- agility
- availability
- security
- global reach
- range of tools and on-demand services
Cloud models:
- Infrastructure-as-a-Service (IaaS) - VMs, networking, load balancers, firewalls; falls into 3 categories: compute, networking, storage
- Platform-as-a-Service (PaaS) - providing code to run on infrastructure without access to it
- Software-as-a-Service (SaaS) - using ready software, can only access configuration
Disaster recovery
DS strategies - increasing cost and complexity, decreasing recovery time further down:
| Type | Description |
|---|---|
| backup and restore | all data is regularly backed up, f.e. to S3; completely new environment is launched upon disaster recovery event |
| pilot light | key infrastructure are kept running in the cloud, AMIs are prepared to run additional systems |
| warn standby | scaled-down version of full environment; ramp up for disaster recovery event |
| multi site | second full environment running |
Recovery Time Objective (RTO) - time it takes to get systems back up and running to the ideal business state (downtime).
Recovery Point Objective (RPO) - amount of data loss (in terms of time) for a production system during disaster recovery event (difference in time between last backup and when disaster occurred).
Expense types running own data center:
- capitalized expenditure (CapEx) - upfront costs for assests suck as building, servers, and supporting equipment; deprecated over the useful life
- operational expenditure (OpEx) - regular day to day expenses on ongoing connectivity, utility and maintenance
High availability (HA) covers local failures such as a disk, power supply.
Fault tolerance is similar to HA, but with a zero downtime.
PaaS examples - Heroku, Azure Logic Apps, Amazon Elastic Beanstalk