Choosing between object storage providers used to be simple: you used Amazon S3. Today the market is crowded, and the differences that matter — egress fees, data residency, real durability, and how "S3-compatible" a provider actually is — are exactly the ones that don't show up on a pricing page's headline number.
This guide walks through what to compare, how the major providers stack up, and when a smaller European provider is the better call than a hyperscaler.
Object storage keeps data as objects (a file plus its metadata plus a unique key) in a flat namespace called a bucket, reached over an HTTP API rather than a filesystem. That design makes it cheap, effectively infinitely scalable, and ideal for backups, media libraries, log archives, data lakes and static assets — anything you write once and read many times.
Almost every provider speaks the S3 API, the de-facto standard Amazon created. That's what lets tools like aws-cli, rclone, Cyberduck, Veeam and most application SDKs talk to any provider with a few config changes.
"S3-compatible" is a spectrum. Basic GET/PUT works almost everywhere, but multipart uploads, presigned URLs, lifecycle rules, versioning and bucket policies vary. Before you commit, test your actual tools against a provider's endpoint — not just a "hello world" upload.
This is where surprises live. Some providers advertise a low per-GB storage price, then charge steeply every time you read your data back out. If you serve media, run restores or move data between clouds, egress can dwarf your storage cost. Providers with zero egress fees are dramatically cheaper for read-heavy workloads — always model your real read/write ratio, not just stored volume.
Where your bytes physically sit is a legal question, not just a latency one. For EU businesses, storing customer data with a provider that keeps it inside the EU under GDPR — rather than on infrastructure subject to foreign data-access laws — is often a hard requirement. If compliance matters, filter for it first; it eliminates most of the list.
Look for how many copies of an object are kept and across how many failure domains. "Eleven nines" of durability is the marketing standard, but what matters is whether data is replicated across drives, nodes or sites, and whether that's included or an upsell.
Compare the whole model, not the storage rate: storage per GB, egress per GB, API request costs (per 1,000 GET/PUT), and minimum object retention. A "cheap" provider with per-request fees can lose to a flat-rate one once you count millions of small objects.
Throughput, request latency and how close the region is to your compute all affect real-world speed. If your servers are in Europe, storage in Europe means lower latency and no cross-continent transfer.
Because everyone speaks S3, switching is easier than with most cloud services — but check for a real migration path, human support, and whether you can pull all your data out without an egress penalty.
If you're an EU business — or serve EU customers — and you want predictable costs, a hyperscaler is often the wrong default. A managed, S3-compatible provider hosted in Europe gives you the same API and tooling, data that stays in the EU under GDPR, and pricing without egress surprises.
That's exactly the gap Internetport's object storage fills: S3-compatible storage in our own European data center, no egress fees, GDPR-compliant, and a free 10 GB tier so you can test your real tools before committing a single krona. It works with aws-cli, rclone, Cyberduck and any S3 SDK out of the box.
Before you sign up with any object storage provider, confirm:
Get those six answers and the right provider usually picks itself. If EU residency and no egress fees are on your list, try our object storage free with 10 GB and benchmark it against whatever you're using today.