Azure Storage redundancy options: LRS, ZRS, GRS, GZRS, and read-access variants
Slide deck explaining Azure Storage redundancy options: LRS, ZRS, GRS, GZRS, and read-access variants (RA-GRS, RA-GZRS), covering copy locations, failure scope protection, and selection guidance.
Azure Storage redundancy options: LRS, ZRS, GRS, GZRS, and read-access variants
Introduction to Azure Storage redundancy options, covering different levels of data protection and availability.
Azure Storage redundancy options: LRS, ZRS, GRS, GZRS, and read-access variants
Introduction to Azure Storage redundancy options, covering different levels of data protection and availability.
Why redundancy?
Redundancy keeps your data available and durable when infrastructure fails. Hardware failures happen (disk/server/rack). Outages can be larger (datacenter, zone, region). Redundancy equals multiple copies of data. Goal: durability plus availability.
Redundancy = copies + location
The main difference between options is where Azure stores the extra copies. Multiple copies of the same data. Copies can be in 1 datacenter, across zones, or across regions. Pick based on the failure scope you want to survive. Redundancy does not equal backup.
Region vs Availability Zone (AZ)
Zones spread risk inside a region; geo options spread risk across regions. Region equals geographic area with datacenters. Availability Zone (AZ) equals separate physical location within a region. Zone outage does not equal regional outage. Not all regions support zones.
Locally Redundant Storage (LRS)
LRS keeps copies in one datacenter: cheap, protects from local hardware failures. Copies within a single datacenter (primary region). Protects against local hardware failures. Does not cover a full datacenter outage. Common fit: dev/test, non-critical data.
Zone-Redundant Storage (ZRS)
ZRS stores copies across availability zones in one region to survive a zone outage. Copies across multiple Availability Zones (AZs). Same region (not geo). Survives a zone outage. Requires zone support (region plus service).
Geo-Redundant Storage (GRS)
GRS adds a secondary-region copy for regional disaster recovery (typically asynchronous). Replicates to a secondary region. Regional disaster coverage. Asynchronous replication (secondary can lag). Secondary used via failover (not normal reads/writes).
Geo-Zone-Redundant Storage (GZRS)
GZRS combines zone resilience in the primary region with geo replication to a secondary region. ZRS in the primary region (zone outage coverage). Geo replication to a secondary region (regional coverage). Asynchronous replication to secondary. 'Zone plus region' protection.
Read-access geo options
RA-GRS / RA-GZRS let you read from the secondary region, but it's read-only and may lag. Read-Access Geo-Redundant Storage (RA-GRS). Read-Access Geo-Zone-Redundant Storage (RA-GZRS). Read-only secondary endpoint (where supported). Eventually consistent (secondary can lag).
Redundancy ≠ backup
Redundancy copies your data; backup lets you restore older versions after mistakes. Redundancy: survive infrastructure failures. Backup/restore: recover from deletion or corruption. Deletion can replicate across all copies. Use data protection features for history/restore.
Redundancy choice checklist
Pick based on the biggest failure you must survive: hardware, zone, or region. Hardware failures only → Locally Redundant Storage (LRS). Zone outage in-region → Zone-Redundant Storage (ZRS). Regional disaster copy → Geo-Redundant Storage (GRS). Zone plus region coverage → Geo-Zone-Redundant Storage (GZRS). Need secondary reads → RA-GRS / RA-GZRS (read-only, may lag).