One of the characteristics of vRanger is that while it’s easy to use, it also provides many options not available in other virtualization-focused data protection solutions. Deduplication is an example of where options abound. vRanger backups can be reduced with patented Active Block Mapping to skip white space, which reduces backup file sizes by up to 1/3 without placing excessive resource load on the backup or production environments. From there vRanger backups can be streamed to any of the leading deduplication appliances or to Quest NetVault SmartDisk – vRanger Edition. We’ve recently gone a step further with EMC’s Data Domain deduplication storage system by integrating vRanger with DD Boost software. Here we’ll explore the mechanics behind the integration, then walk through the setup.
Data Domain provides high speed, inline deduplication and storage of vRanger backups, and can replicate the backups over the WAN to other Data Domain systems for enhanced disaster recovery purposes. For more detailed information on Data Domain, please visit EMC’s Data Domain site. Historically, many administrators stream vRanger backups to Data Domain systems using CIFS repositories.
Data Domain also provides backup application integration through DD Boost, which has been included in vRanger since version 5.4. There are two components to DD Boost:
- The DD Boost server that runs on the Data Domain system or gateway, which requires a license key to operate that can be obtained from EMC
- The DD Boost libraries that ship with vRanger, which is enabled out of the box and does not require additional licenses
DD Boost enables vRanger to participate in the deduplication process through distributed segment processing. Before streaming the backup copies, vRanger communicates via the DD Boost libraries with the DD Boost server on the Data Domain system to determine which blocks are redundant (i.e., are already resident on the Data Domain system) and which blocks are unique. Once that’s determined, vRanger streams only the unique blocks to the data domain system. While vRanger participates in the deduplication process using DD Boost, there’s no additional load placed on the host or on vRanger with distributed segment processing. The DD Boost libraries are stateless, and you won’t see a spike in CPU or RAM when using DD Boost during the backup process.
When using DD Boost, vRanger backups are streamed to a DD Boost repository instead of a CIFS repository. With the DD Boost repository, the Data Domain system exposes disk volumes called Storage Units to vRanger. You can either create new Storage Units or use existing ones when configuring a DD Boost repository. When using existing Storage Units, vRanger will not over-write existing data. In fact, multiple vRanger servers can share the same Storage Units on a Data Domain system.
All vRanger backup architecture options can use DD Boost, including vRanger server-based backups (with vRanger installed in a virtual machine or on a physical server) and VA-based backups.
Higher throughput can be achieved by sending only the unique blocks to the data domain system compared to sending all the blocks, but results may vary depending on the type and rate of data change and the workload on vRanger. In general, higher throughput can be achieved with higher data redundancy and higher vRanger workloads. Customers are reporting on average a 50% reduction in backup times using DD Boost compared to CIFS, with some customers reporting as much as a 75% reduction in job times.
Setting up vRanger to use DD Boost is a 2-step process. First, configure a DD Boost repository. Right- clicking anywhere in the left hand pane in the My Repositories view will launch the repository wizard. Alternatively, you can also click the “Add” icon at the top of the left hand pane in the My Repository view to launch the repository wizard. If you’ve ever created a repository in vRanger, creating a DD Boost repository will look familiar. The one exception is that we need to create or use an existing Storage Unit. One repository needs one storage unit. In our lab you can see we had several existing Storage Units:
Second, configure a backup job to stream to the DD Boost repository. Launch the backup job wizard to create a new backup job, or edit an existing job to use the DD Boost repository. When you’re at the step in the backup wizard to choose the repository, select the DD Boost repository we just created:
Now you can start enjoying faster backups.
For more detailed information on vRanger integration with EMC Data Domain Boost, please see: