When would you use object storage? The short answer is when you need to scale.
A typical file system requires a single writer to ensure block mapping, directory entries, and partial file writings are performed atomically. Object storage does not have directories, but only allows you to add or replace objects. Having these limitations makes it easier to distribute your object storage across multiple servers.
AWS S3, Azure BLOB, and Google Cloud Storage are the most popular object storage services. Let's check out their different tiers.
🟢 AWS S3
- ➡️ Amazon S3 Standard
High throughput and low latency make this the best storage choice for information that is accessed often, such as dynamic websites, cloud applications, content distribution, data analytics, and gaming.
- ➡️ Amazon S3 Standard—IA
A storage option for data that is accessed less often, such as disaster recovery.
- ➡️ Amazon Glacier
A highly durable storage system optimized for "cold" data, such as end-of-lifecycle records that are kept for compliance and regulatory backup purposes. It is immutable and encrypted for long-term storage.
🟢 AZURE BLOB
- ➡️ Hot access tier
For data in active use or about to be used and its subsequent migration to the Cool storage tier.
- ➡️ Cool access tier
For long-term storage of data, including disaster recovery datasets, short-term backups, older media content and large data sets.
- ➡️ Archive access tier
For archived data that can tolerate hours of retrieval latency and is kept longer than 180 days.
🟢 GOOGLE CLOUD STORAGE
- ➡️ Hot (high-performance) storage
Regional and multi-regional high-frequency access information storage system.
- ➡️ Multi-regional storage
Stores information that is often accessed around the world, including videos, website content, and mobile games.
- ➡️ Regional storage
Provides frequent access to information in the corresponding region.
- ➡️ Nearline (cool) storage
Suitable for backups and long-tail multimedia content that only needs to be accessed occasionally.
- ➡️ Coldline (cool) storage
Archival data and disaster recovery. Suitable for data accessed less than once a year. Now, the question of price – generally, the more performant and redundant the storage class, the more expensive it is.
💡 If you're looking for a hybrid solution, you should definitely consider Minio. It implements the same public API as Amazon S3, so applications that talk to Amazon S3 can also talk to Minio. Minio is an object store that can store unstructured data such as photos, videos, logs, backups, and container/VM images. Files are separated into “buckets”, which are logical separators for your data, which need to be passed to applications along with connection keys, secret keys, and HTTP addresses.
Say hi at hello@devolut.io and get more out of your cloud provider for less.