Azure Storage: Design for Efficient Usage with Access and Performance Tiers

The final key configuration we must understand in Azure Storage is its access and performance, tier models. Azure Storage offers different access tiers, allowing you to store blob object data in the most cost‑effective manner. Available access tiers include:

Hot Tier: Optimized for storing data that is accessed frequently. The hot access tier has higher storage costs than cold and archive tiers, but the lowest access costs. Example usage scenarios for the hot access tier include data that's in active use or is expected to be read from and written to frequently, data that's staged for processing, and eventual migration to the cold access tier. A classic example would be the videos for Youtube. We serve them daily, so we want fast access and good performance.

Cold Tier: Optimized for storing data that is infrequently accessed and stored for at least 30 days. The cold access tier has lower storage costs and higher access costs compared to hot storage. This tier is intended for data that will remain in this tier for at least 30 days. Example usage scenarios would be short‑term backup and disaster recovery, older data not used frequently, but expected to be available immediately when accessed, and large datasets that need to be stored cost-effectively while more data is being gathered for future processing. The classic example would be the log data from Youtube. We store it in case there is a need, but we don't care about having a slower performance since it is not mission‑critical for the performance of the site.

Archive Tier: Optimized for storing data that is rarely accessed and stored for at least 180 days with flexible latency requirements on the order of hours. The Archive access tier has the lowest storage costs, but higher data retrieval costs compared to hot and cold. Data must remain in the archive tier for at least 180 days or be subject to an early deletion charge. Data in the archive tier may take several hours to retrieve, depending on the specified rehydration priority. While a blob is in archive storage, the blob data is offline and can't be read or modified. Example usage scenarios for the archive tier include long‑term backup, secondary backup, and archival datasets, original (raw) data that must be preserved, even after it has been processed into a final usable form, and compliance and archival data that needs to be stored for a long time and is hardly ever accessed. The classical example would be audit data we don't need in the regular day‑to‑day operations, but we may need it in case we must pass a compliance case.

Additionally to the access tiers, given its wide usage, blob storage also has a premium tier on performance for high throughput applications. Premium performance storage is ideal for workloads that require fast and consistent response times. It's best for workloads that perform many small transactions like interactive workloads. For example, in an e‑commerce application, less frequently used items are likely not cached. However, they must be instantly displayed to the customer on demand. Analytics is another good case. In an IoT scenario, lots of smaller write operations might be pushed to the cloud every second. Large amounts of data might be taken in, aggregated for analysis purposes, and then deleted immediately. The high ingestion capabilities of premium block blob storage make it efficient for this type of workload. Artificial intelligence is another good use case since it deals with the consumption and processing of different data types like visuals, speech, and text. Finally, data transformation, which includes processes that require constant editing, modification, and conversion of data and require instant updates. It is important to mention that Azure Data Lake Storage Gen2 is not supportive of the premium tiers, as of April 2021. All of these options make storage accounts really flexible in terms of how we want to access our data and which price to pay for it.