01 - Cloud scope of Application
Selecting regions
We should select region based on several aspects:
- We need to look if we already have resources in some region.
- Some Azure regions doesn´t support all things so we need to inform ourselves about regions.
- Then we need to look if application/customer requires us to store data in some region/country.
- And last but not least we need to consider physical distance between region and customer. (This can be partially mitigated by using Azure CDN.)
Single region without redundancy
We will be prune to any outages that could happen inside Azure warehouses. But on other hand this is the easiest and cheapest solution.
Single region with zone redundancy
Some regions, not all of them, supports Zone redundancy with three seperated zones. We are still prune to region-wide issues connected to Azure updates, but we are protected against outages.
Each zone have different:
- Power source and backup system
- Air cooling system
- Internet connectivity
- Protection against natural disaster
Single region with fallback region
We recognize three seperate fallback setups. First is backup setup, it means that all our data are stored also in backup region, but we have no way how to actively use them from this region.
Second is read-only fallback, it means that we can read data from second region. So for instance we are able to keep websites online and just don´t allow any changes.
Third is active fallback, that in case that main region drops out is able to replace entire functionality.
Multiple regions setup
Multiple regions setup is one of the hardest thing to do and I would not recommend it to do it without devops engineers.
You need to carefully manage data synchronization, replication logic and many other aspects.