Deduplication and VMware Backup Sprawl
by Mike Beevor and Ricky El-Qasem
In this white paper you can learn more about Veeam Deduplication and Compression and VMware Backup Sprawl problems. Find out how Veeam makes a positive impact on storage environments, as well as best practices to maximize your VMware Backup deduplication.
The whitepaper also provides some comments and best practices on how to best maximize the officiency of your VMware backup processes with respect to deduplication.
VM Sprawl Control and Prevention
Virtualization makes it much easier to create virtualized servers and that is where the virtual machine (VM) sprawl problem comes in. Over time, the virtualized infrastructure overflows with VM that are not really needed and have been created just because it is easy to do so. With VMs being created for every use possible, licensing and change management documents and processes are stretched to their limits, new hosts are being provisioned and resource pools are being used up. With that many VMs around, virtual machine sprawl management becomes difficult, and tracking them for disk space and resource usage is an IT administrator’s last concern. This might eat away at the benefits you get through virtualization.
Some of the things you can do to control VM sprawl is implement a formal request process that calls for an explanation why a new VM is needed. This should cut down on the creation of VMs that are truly not required. You can also limit the number of people who are tasked with creating a virtual machine and have that person delete virtual machines that are no longer in use.
There are certain virtual machine sprawl tools that can help with virtualization sprawl. They assign all of the VMs in the datacenter. All other VMs and resources that are being used are immediately shown in a dashboard. These tools help identify abandoned VMs and minimize unnecessary license costs and resource burdens.
About the authors
Whilst deduplication, as a concept is relatively straightforward, merely being the removal of identical blocks from a location, and a marker left in place indicating the position of the block itself when restored to full size, there are many methods by which this can be achieved. In an examination of two major types of deduplication, it's possible to highlight some of the differences and their impact on the deduplication process.
Deduplication at the storage layer has become an almost commodity feature in today's market with DataDomain widely accepted as leading the market in the process. Exagrid, HP, Greenbytes and several other vendors also provide solid solutions for this. Deduplication at the storage layer is always performed as a global operation at the device level, with expansions into offering global deduplication across many devices in a storage environment.
The ability to deduplicate in a virtualised environment is less common, and to find the ability in a VMware Backup product is rarer still. Veeam were one of the first vendors to introduce inline block level deduplication within its products, making the functionality available in the v1.0 release. It has taken up until now for the concept of a back up product providing deduplication to become thought of as necessary, and the move towards commoditising the feature is gathering speed.
This isn't to imply that the chosen methods of deduplication are identical. The majority of vendors delivering this feature are offering it as "Global Deduplication" and will house a single deduplication store per installation, shared between all back up jobs. All unique blocks across all included Virtual Machines are stored. Whether or not this is the most appropriate method for a virtualised environment back up is open to some discussion, but there are some significant drawbacks that I've observed.