What is Veeam backup to object storage?
Veeam backup to object storage means using an S3 bucket, in a public cloud or on premises, as a repository where Veeam Backup & Replication keeps backups. One bucket can play one of four roles, and its location decides restore speed and cost.
What four roles can an S3 bucket play?
Veeam can treat object storage as a plain destination or as a stage in a longer lifecycle. The role is chosen when the repository is added, and it determines when data lands in the bucket.
| Role | What lands in the bucket | Typical use |
|---|---|---|
| Direct to object storage | Backups written straight from the job | A single repository with nothing in front of it |
| Performance tier | Each new restore point, in a scale-out repository | Recent restore points kept for the fastest restores |
| Capacity tier | Older restore points moved or copied off the performance tier | Longer retention without growing primary disk |
| Archive tier | Long-retention points moved from the capacity tier | Rarely read data on cold, low-cost storage classes |
In a scale-out repository, the capacity tier can run in copy mode, which duplicates every restore point as it is created, or in move mode, which offloads points once they pass an operational restore window. Veeam writes data as many small blocks and records which blocks belong to which restore point, so a new full backup can reuse blocks already in the bucket instead of storing a second complete copy. This block reuse is also why locks and retention behave differently from file-based repositories, as explained under Veeam immutable backup.
Why do storage-side lifecycle rules break the chain?
Object storage normally offers lifecycle rules that expire old objects or move them to cheaper classes after a set number of days. On a bucket that Veeam writes to, those rules cause damage. Veeam tracks every object and decides for itself when a restore point has aged out. An object that a storage rule deletes or moves is still referenced by newer restore points because of block reuse, and the next restore finds a missing piece. Retention belongs to Veeam alone, and a bucket holding Veeam data carries no expiration or transition rules.
The same discipline applies to retention policy design: one system decides when data expires, and the other enforces the lock.
Local bucket or cloud bucket?
Where the bucket lives sets the cost of a restore. A bucket on the local network returns data at LAN speed, with no per-gigabyte retrieval fee. A bucket in a public cloud is off-site by default, but a large restore is limited by the internet link and may carry egress fees and request charges that vary by provider. Many designs keep recent restore points local and offload older ones to the cloud, which makes the cloud a place data rarely has to return from.
A word of caution about the label offsite. A bucket in the same room as the production racks, or a cloud region reached only from the same compromised network, is a second copy in the same failure domain. Distance and separate credentials are what make a copy off-site.
What Object Lock adds
For immutability, the bucket has to be created with versioning and S3 Object Lock enabled. Veeam then sets the lock dates and the storage enforces them. Veeam extends each lock by a block generation period, which adds days to a few weeks depending on the object storage type, so that blocks shared by older and newer restore points remain protected. Capacity planning has to include that extra locked data.
Frequently asked questions
Can any S3-compatible storage be used with Veeam?
Only systems that Veeam supports or that have been validated for it, since Veeam relies on specific API behavior. Object Lock support has to be confirmed separately for immutability.
Does object storage make Veeam backups faster?
Not by itself. Speed depends on the proxies, the network and the object storage's throughput. The change object storage brings is capacity that scales out and immutability through Object Lock.
Can the bucket be shared between several Veeam servers?
Usually each repository uses its own bucket or folder, because Veeam manages the contents and one server would not recognize another's objects as its own.
