You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Differences between Availability Set and Availability Zone
Availability Set
Availability Zone
Protect from Hardware failures within data centers
Protect from entire data center failure
SLA 99.95 %
SLA 99.99%
Availability Set
No additional cost
Same data center
At least 2 VMs in set
Availability Zone
3 per region (not all regions supported)
Separate power source, network and cooling
Characteristics of different redundancy options
Redundancy Options Scenarios
Scenario
LRS
ZRS
GRS/RA-GRS
GZRS/RA-GZRS (preview)
Node unavailability within a data center
Yes
Yes
Yes
Yes
An entire data center (zonal or non-zonal) becomes unavailable
No
Yes
Yes
Yes
A region-wide outage
No
No
Yes
Yes
Read access to your data (in a remote, geo-replicated region) in the event of region-wide unavailability
No
No
Yes (with RA-GRS)
Yes (with RA-GZRS)
Designed to provide __ durability of objects over a given year
at least 99.999999999% (11 9's)
at least 99.9999999999% (12 9's)
at least 99.99999999999999% (16 9's)
at least 99.99999999999999% (16 9's)
Supported storage account types
GPv2, GPv1, Blob
GPv2
GPv2, GPv1, Blob
GPv2
Availability SLA for read requests
At least 99.9% (99% for cool access tier)
At least 99.9% (99% for cool access tier)
At least 99.9% (99% for cool access tier) for GRSAt least 99.99% (99.9% for cool access tier) for RA-GRS
At least 99.9% (99% for cool access tier) for GZRSAt least 99.99% (99.9% for cool access tier) for RA-GZRS
Availability SLA for write requests
At least 99.9% (99% for cool access tier)
At least 99.9% (99% for cool access tier)
At least 99.9% (99% for cool access tier)
At least 99.9% (99% for cool access tier)
Redundancy Options Simple
locally redundant storage
Designed to provide at least 99.999999999% (11 9’s) durability of objects over a given year by keeping multiple copies of your data in one data centre.
zone redundant storage
Designed to provide at least 99.9999999999% (12 9’s) durability of objects over a given year by keeping multiple copies of your data across multiple data centres or across regions.
geographically redundant storage
Designed to provide at least 99.99999999999999% (16 9’s) durability of objects over a given year by keeping multiple copies of the data in one region, and asynchronously replicating to a second region.
read-access geographically redundant storage
Designed for to provide at least 99.99999999999999% (16 9’s) durability of objects over a given year and 99.99% read availability by allowing read access from the second region used for GRS.