Why Partitioning Matters
Azure Storage distributes data across partition servers to scale horizontally. Each partition server has throughput limits. If your data or access patterns concentrate on a single partition ("hot partition"), you hit throttling — even if the overall storage account has capacity. The key to high performance is spreading load evenly across partitions.🗃️ Table Storage
Table Storage gives you explicit control over the PartitionKey — making it the most impactful design decision. Simulate date-based, category-based, unique, composite, and hash-bucket strategies.
- Visualize entity distribution across partitions
- Detect hot partitions exceeding 2,000 ops/sec
- Compare 6 different PartitionKey strategies
- Factor in query patterns and data skew
📦 Blob Storage
Blob partitioning is driven by naming patterns. Sequential names are the #1 anti-pattern. Simulate how different naming conventions (hash-prefix, GUID, timestamp) affect partition distribution.
- Compare 6 blob naming conventions
- Visualize partition server distribution
- Detect hot-spot risk from sequential names
- Model container strategies and access patterns
📁 Azure Files
Azure Files performance is driven by share size, tier, and protocol rather than naming. Simulate IOPS, throughput, and capacity against provisioned limits across shares.
- Model Premium, Hot, and Cool tier limits
- Visualize IOPS and throughput gauges per share
- Plan multi-share distribution strategies
- SMB multichannel and NFS recommendations
Quick Decision Guide
| If you need… | Start with… |
|---|---|
| High-throughput structured data, batch operations | Table Storage simulator — PartitionKey design is critical |
| Large-scale blob ingestion (IoT, logs, media) | Blob Storage simulator — naming pattern determines scale |
| Shared file system (SMB/NFS lift-and-shift) | Azure Files simulator — tier and share sizing are key |