There are some important concepts about infrastructure layout in OpenStack.
Regions: invented by
AWS. Region is based on the geography location. It is more like a physical datacentre.
Regions are isolated between each other. But regions can share the keystone and
dashboard. Users can choose a close region to deploy their application. Region
is also considered in DR plan.
Availability Zone(AZ): AZ is the division in a Region. Usually
when the hardware is deployed, separated
infrastructure such as power supply, air-conditioning or network connections. It
provides some kind of redundancy capacity. AZ is visable to the user and end
user need to select an AZ when they create an instance.
Cell: cell is used in OpenStack for scaling and
expanding. In large scale of OpenStack
deployment, DB and AMQP will be the bottleneck, in that case, cell can be introduced,
each cell can own the DB and AMQP.
Host Aggregate: a
bounch of computing nodes and usually these nodes have same character such as
SSD disk, same OS. Host Aggregation is used by OpenStack internally (manually
by scheduler). Host Aggregate usually won’t be visible to End users
