Why Redundancy Matters
Data is the backbone of every workload. If it’s lost or unavailable, businesses can face downtime, compliance issues, or financial loss.
Azure Storage provides built-in redundancy options to ensure durability and availability across failures — from a single disk crash to a complete regional outage.
As a Solution Architect, your job is to choose the right redundancy level based on business continuity requirements, cost, and compliance.
Azure Storage Redundancy Options
1. Locally Redundant Storage (LRS)
-
Keeps 3 copies of data within one datacenter.
-
Protects against hardware failure (disk/server).
-
Lowest cost redundancy.
-
Does not protect against datacenter outages.
Best for: Non-critical data, dev/test workloads.
2. Zone-Redundant Storage (ZRS)
-
Keeps 3 copies across 3 availability zones within a region.
-
Protects against entire datacenter failure.
-
Higher availability than LRS (supports 99.99% SLA).
-
More expensive than LRS.
Best for: Production data that requires protection from datacenter failures.
3. Geo-Redundant Storage (GRS)
-
Replicates data asynchronously to a secondary region (hundreds of km away).
-
Primary region uses LRS, plus 3 copies in secondary region.
-
Total of 6 copies (3 primary + 3 secondary).
-
Failover must be initiated by Microsoft.
Best for: Disaster recovery compliance where data must be recoverable in another region.
4. Read-Access Geo-Redundant Storage (RA-GRS)
-
Same as GRS but with read access to the secondary region.
-
Allows apps to read data from secondary endpoint even during primary outage.
Best for: Apps that require cross-region read redundancy (e.g., global reporting).
Performance Tiers
-
Standard (HDD-based) → Cost-effective, good for general-purpose workloads.
-
Premium (SSD-based) → Low latency, high IOPS, designed for I/O-intensive apps like databases or VM disks.
Blob-specific tiers:
-
Hot → Optimized for frequently accessed data.
-
Cool → Lower cost, higher access fees (good for infrequent access).
-
Archive → Cheapest storage, but requires rehydration (hours) before access.
Example Enterprise Scenario
A healthcare provider must:
-
Store patient records with 99.99% uptime within their region.
-
Maintain long-term backups that must survive a regional disaster.
-
Keep older data (7+ years) at lowest possible cost but retrievable.
Correct design:
-
Use ZRS for primary patient data storage.
-
Use GRS for disaster recovery backup.
-
Store older records in Archive tier.
Confusion Buster
-
LRS = same datacenter only.
-
ZRS = multiple datacenters in the same region.
-
GRS = cross-region, secondary not normally readable.
-
RA-GRS = cross-region with read access.
-
Standard vs Premium = based on performance needs, not redundancy.
Exam Tips
-
“Cheapest redundancy option?” → LRS.
-
“Protect against datacenter failure within a region?” → ZRS.
-
“Protect against region failure?” → GRS.
-
“Cross-region read access?” → RA-GRS.
-
“High IOPS, low latency disks?” → Premium SSD.
-
“Which blob tier for archival data?” → Archive.
What to Expect in the Exam
-
Direct Q: “Which redundancy keeps 3 copies across multiple availability zones?” → ZRS.
-
Scenario Q: “Company requires cross-region redundancy with read access to secondary region.” → RA-GRS.
-
Scenario Q: “Which storage tier for frequently accessed production logs?” → Hot.
-
Trick Q: “GRS allows read access to secondary.” → False (only RA-GRS does).