What is Veeam backup to object storage?
Veeam backup to object storage is the use of an S3 bucket as a place where Veeam Backup & Replication keeps backups. The bucket can sit in a public cloud or on S3-compatible storage in the organization's own data center.
To Veeam Backup & Replication, the bucket is one more repository. What changes is how the backups are stored, who can reach them and how fast they come back.
A bucket instead of a backup folder
A classic Veeam repository is a folder on a Windows or Linux server. Anyone with administrator rights on that server can open the folder and delete what is inside. When the disks fill up, the team trims retention or buys a bigger server.
A bucket works more like a storage unit with a keypad. Veeam reaches it only through the S3 interface, using keys that can be limited to that one bucket. It grows by adding capacity rather than replacing a box.
A bucket can also lock each restore point until a set date with S3 Object Lock, which a plain file share cannot do on its own. This is the basis of Veeam immutable backup on object storage.
Where the bucket fits in a Veeam setup
Veeam can give a bucket one of four jobs. The choice ties into how they already split backup storage tiers.
| Role | What the bucket does | Typical use |
|---|---|---|
| Direct target | Backup jobs write straight to it | Simple setups with one main repository |
| Performance tier | Receives each night's new restore points first | Object storage on premises as the primary copy |
| Capacity tier (Cloud Tier) | Gets older or copied restore points from local disk by policy | Longer history, or a second copy |
| Archive tier | Holds rarely read, long-term restore points | Compliance and yearly copies |
Inside the bucket, Veeam stores backups as many small blocks. A new synthetic full can reuse blocks already there, so a weekly full adds far less capacity than its size suggests.
Choosing between a local and a cloud bucket
A cloud bucket is offsite with no hardware to buy. On restore day, though, every terabyte crosses the internet link, and many providers charge for data read out. A large restore can take days. A bucket on the local network returns data at LAN speed with no retrieval fee, which is what drives restore throughput in a real incident.
Common mistakes with either option:
- Using Move without Copy. Moving old restore points to the bucket stretches history but leaves only one copy of them.
- Adding bucket lifecycle rules. Veeam runs retention itself, and its documentation warns that lifecycle rules on its buckets can break backups and restores.
- Giving the backup server full-control keys. A stolen Veeam server with broad keys can do far more damage than one limited to its own bucket.
- Calling a bucket offsite when it sits in the same room. A fire or flood takes both copies.
ARTESCA and Veeam backup to object storage
ARTESCA is S3 object storage that runs on premises, so it can take any of the four bucket roles at the main site or a second one. Veeam Ready lists it as an object repository with immutability and SOSAPI support, the interface that reports capacity back to Veeam. Scality's Veeam compatibility page names Veeam Backup & Replication, Veeam Backup for Microsoft 365 and Kasten by Veeam as validated.
Restores from ARTESCA travel at local network speed with no per-gigabyte charge. It scales to a validated 8.5 PB of object storage. Veeam, not ARTESCA, decides when restore points expire.
Related terms
- Veeam Backup & Replication: the backup application that writes to these buckets.
- Veeam immutable backup: how Veeam locks restore points until a set date.
- Backup storage tiers: placing backup copies on storage of different cost and speed.
- Restore throughput: how quickly backup data can be read back.