Reputation: 394
Is Amazon S3 an object store or a key-value store? I am getting conflicted answers on the web when Googling the question. On one hand, I understand that the objects that you upload to S3 are binary object data (not data blocks), which gets overwritten fully when you update them. On the other hand, you can also reference objects in S3 using a key, which has a prefix and object name, and a value, which is the object itself.
So, what type of storage is S3? Object store or key-value store?
Upvotes: 2
Views: 1997
Reputation: 1298
When you store key-value, key is generally string and value could be anything i.e. string or even binary object. If you consider path of file/object stored in S3 bucket as a key and the file/object itself as value. By this logic you can say, S3 stores is key-value pair.
However, S3 is object store which stores the paths as just key and not as real folder structure as Azure ADLS Gen2 Hierarchical File System does. You can access files under a specific path in S3. But there is know Hierarchical File System under the hood.
Hope this answers your question.
Upvotes: 0
Reputation: 23602
S3 is not for storing key-value pairs. That is more for DynamoDB which is AWS's key-value database offering.
You access objects within S3 using the object 'key' but at a conceptual level, you can look at this as the filename.
S3 is for object storage, it is not a key-value store in the typical meaning of the phrase.
Upvotes: 3