Object Storage

Object Storage lets you:

• Host project files and control access by making them public or private.
• Store large volumes of archived data (up to 5 TB per file) with limited availability.
• Collaborate on projects using data inside a distributed organization.

Physical media is located in highly secure facilities in the Nebius data center.

Data is stored in an encrypted format and no one with access to the physical media can read the data.

By default, storage is accessed via HTTPS.

Object Storage has internal SLAs and SLOs that guarantee a specific data transfer speed for the customer, which depend on the storage class (among other things).

The Object Storage response time depends on multiple factors:

• Client-side performance (network speed, CPU load, or disk subsystem load).
• The speed of trunk connections, connections between data centers, or client-server connections.
• Performance of Object Storage itself.

As a result, we cannot specify a particular response time. We do not consider a sharp increase in response time or a noticeable decrease in the data transfer speed to be the norm, and we are constantly working to improve the technical characteristics of Object Storage.

You can create and manage buckets using the Nebius management console. The Object Storage API is also compatible with the Amazon S3 API, so you manage objects using tools designed for S3.

To start using the S3 API:

• Create a service account.
• Assign a role to the service account.
• Create a static access key.

Static key authorization is required to access the HTTP API directly and is supported by the available tools. For more information about working with the S3 API, see the corresponding documentation.

You can store data in any format. Object Storage saves data in its original form without any modification.

When storing data in Object Storage:

You get unlimited disk space.

In Compute Cloud, you are limited to a specified disk size.

Data is always available and can be managed directly.

Managing data on a Compute Cloud disk means managing a VM that the disk is attached to. Data on the disk is only available when the VM is running.

Data is easily shared: file links with limited lifetimes.

To share data stored on a Compute Cloud disk, you need to grant a user access to the VM over SSH or RDP. You also have to configure access to specific files yourself.

You pay for data storage, operations, and outgoing traffic.

When you store your data on a Compute Cloud disk, you pay for the disk, VM resources, network resources, and outgoing traffic.