Don’t let the number of files or the total size limit you. portvMind Object Storage works directly with your existing tools and libraries through its S3-compatible API; you store everything from backups to static web content under a single bucket model, with no need to set up a separate storage layer.
Object Storage is a storage model that keeps files as objects rather than as blocks or a file system; each object has an ID, data and metadata. Access is over HTTP, through API calls.
What sets Object Storage apart from block storage is its purpose. You don’t keep a database’s data files in object storage — that needs block storage. But keeping a database’s nightly backup, your application’s user uploads, video files or log records you must retain for six months in object storage is both more economical and more scalable.
Thanks to the S3-compatible API, SDKs and tools written for AWS S3 — backup software, CDN integrations — can work without extra development. Migration is usually just an endpoint change.
Use your existing S3 clients and SDKs without rewriting them; in most cases an endpoint change is enough.
No capacity planning required; you grow without performance loss as the number of files increases.
Permission management per bucket and object, time-limited signed access links.
Keep earlier versions of the same file and roll back when needed.
Define policies for automatic deletion or archiving after a set period.
Data transfer is not charged for access from your servers in the same region.
| Feature | Detail |
|---|---|
| API | S3-compatible |
| Access control | Permissions per bucket and object; time-limited signed links |
| Versioning | Supported |
| Lifecycle rules | Automatic deletion and archiving policies |
| Encryption | Encryption at rest, TLS in transit |
| Metering | Per minute — based on stored volume and request count |
Object size limits and the durability model are covered in the documentation.
For detailed technical specifications and release notes, see Documentation
Enter the data volume you want to store and see the total cost instantly.
Metering: Per minute — based on stored data volume and request count
Data transfer is not charged for access from your resources in the same region.
Your resources are metered for as long as they are running — you never pay for time you don’t use. In the console you can define a monthly or period-specific budget and get alerts when your spending reaches the 50%, 80% and 100% thresholds. Turn on overrun forecasting to see in advance whether your current consumption rate will take you over budget by the end of the period.
Alerts arrive by e-mail; with webhook support they can also be sent to Slack, Microsoft Teams or your own application.
Budget alerts send notifications; they do not stop resources automatically. You decide which resource to scale down.
The full pricing modelIn model training, data volume grows quickly: the raw dataset, preprocessed versions, the checkpoint files of every experiment and the model artifacts taken to production. Keeping them on Block Storage is both expensive and unnecessary; this data is accessed sequentially and in bulk, not randomly.
Object Storage is the right layer for this job. You write your dataset through the S3-compatible API and read it directly from your GPU nodes. Because it is within the same region, this traffic is not charged — reading the dataset many times during training creates no extra cost.
Versioning can be used for checkpoint files and lifecycle rules for cleaning up old experiments; your experiment archive manages itself.
Your training data is often your organization’s most sensitive data. Every object kept in Object Storage stays in the Istanbul or Ankara data center; you can document with audit records that the data did not leave the country during model training.
In most cases you only need to change the endpoint address; authentication and the request format are compatible with the S3 standard.
Yes, you can set access permissions as public or private per bucket or object.
Based on the total stored volume and outbound traffic. Internal traffic within the same region is not charged.
Yes. Your GPU nodes access Object Storage in the same region directly and this traffic is not charged. Reading the dataset many times during training creates no extra cost.