Late last year Microsoft released Azure Data Box Next Gen, and, for US, EU, UK, Canada and US Gov, this version has fully replaced the old generation. Other region roll-out is still in progress. Using Azure Data Box Next Gen is supported with Veeam and is significantly faster than the previous generation.
We often get questions about seeding data with Veeam Backup and Replication. So, I wanted to cover briefly how that works.
1. Ordering Azure Data Box
- Order your Azure Data Box with ‘Import to Azure’. Selecting the appropriate subscription and resource group.
- Select source country, associated with the address this Data Box will be delivered to, select the destination Azure region, and choose from the Data Box 120 or Data Box 525 as they have the same form factor. Pricing of course will be different. There is a service fee, shipping fee and fees for extra days over the built-in 10 or 20 days depending on the Data Box capacity.
- Give your order a name, select a brand-new storage account as a destination (it will be locked during Data Box migration). This storage account cannot be changed afterwards so ensure it has the desired redundancy and features. Keep in mind that Versioning with Immutability, Archive Tier and Veeam Vault are not supported with Azure Data Box.
- You can use customer managed keys, bring your own password or use double encryption if you like.
- Now for the most crucial part, make sure the address and all contact information is correct, and ensure the people on the receiving end know what to expect, so that the delivery can progress smoothly.
2-3. Setup Azure Data Box and Configure As SOBR Capacity Tier
Upon arrival, ensure you connect Azure Data Box to your infrastructure and setup REST API access. After that you can add it to Veeam Backup & Replication. Configuring Azure Data Box as a Capacity Tier extent with “Copy backups to object storage as soon as they’re created” in your Scale-Out Backup Repository (SOBR). Move mode is not recommended.
In our test environment we were quite impressed with the results of Azure Data Box Next Gen. With the 2x100Gbps QSFP ports mean you might get a bit faster in your environment, depending on your source configuration and block size:
| Azure Data Box 80TiB | 300MB/s | ~2.5 Gbps |
| Azure Data Box Next Gen 120TiB | At least 625MB/s | ~5 Gbps |
Veeam Backup and Replication uses the REST API endpoint (Azure Blob) to talk to Azure Data Box, writing objects, so they can have identical properties when they arrive in the destination storage account. We used 1MiB blocks as this is common with Veeam, but can be changed on a per-job level. The REST API is usually not as fast as writing directly via SMB but has been significantly improved in this generation.
4-5. Prepare for Shipping, Ship and Synchronize Data.
When backup data has fully copied to Azure Data Box, you need to put the Azure Data Box tier into maintenance mode in Veeam and prepare the device for shipping, wait for it to ship and then add the destination Storage Account to Veeam Backup and Replication and synchronize the data.
Deciding to Seed or Using Internet
Now, how to decide whether to use your internet or Azure Databox Next Gen:
| Have dedicated 10Gbps internet circuit available? | Use Internet – ~1TiB/13 min |
| Sharing 10Gbps internet circuit with other things? | Use Azure Data Box – ~1TiB/27 min Service fee $350-1,500 Shipping fee $1-200 Data Box 120 - First 10 days onsite are free Data Box 525 – First 20 days onsite are free + Extra day fees $50-200 Service and extra day fees depend on Azure Data Box size (120TiB vs 525TiB). Shipping costs vary depending on the region. |
| Have 1Gbps internet circuit only? | Use Azure Data Box |
| Using Azure Blob with Immutability, Archive Tier or Veeam Vault? | Use Internet |
When you order make sure the desired generation and form factor is available in your region, and remember the device has QSFP on-board, so be ready to have somewhere to plug it in. And make sure you follow up with your facilities team to make sure they know to expect your order!
We’ve also had some previous questions around using Azure Data Box for restoring data from Azure back to on-premises, this is not as typical of a scenario, but keep in mind that Azure Blob API versions are different on Azure Data Box when compared to Azure Storage - the same limitations around Versioning, Immutability and Archive Tier apply.
Another common question is, what if I have several petabytes of data to upload using Azure Data Box. Now with Azure Data Box Next Gen, each can store 525 TiB, with the same form factor as the 120 TiB unit. This would mean the data copy needs to be segmented between SOBRs, storage accounts and multiple Data Boxes.
Hope this was useful! It’s certainly fun to see how much faster the offload process is now versus what we used to see.