OBS vs S3

[outcatcher]

When using the Open Telekom Cloud provider, you may have noticed that there are two sets of resources/data sources for Object Storage: s3_ and obs_.

First of all ― OBS implements an S3 interface allowing any S3 client to use it, including file uploading and downloading. That's basically the main reason why, originally, opentelekomcloud provider is using s3_ resources ― they are mostly a copy-paste from AWS provider and using AWS SDK for Go.

On the other hand, in the documentation, OTC addresses object storage as "OBS", not "S3" ― that can lead to confusion, when a user is searching for OBS in the provider documentation and can find nothing. Also, OBS functionality is not restricted by S3 API, so potentially it can provide functionality different from S3.

So obs_bucket and obs_bucket_object resources were implemented in #467 and become part of the provider in release v1.17.0. obs_ resources are using internal SDK, which is currently a part of OTC SDK for Go.

At the current state of development, obs_ resources' main advantage is the possibility to use them without providing AK/SK to the provider. Simple token-based authorization (user+password or token) is just enough.

In the future, we plan to deprecate and remove s3_ resources, but only when we will be sure that obs_ resources cover everything that s3_ resources do.

---

I hope this article was useful for you and answered questions about s3_ and obs_ duality you might have.