By Infrascale Guest Blog 1

The cloud backup market is certainly a cluttered one. With new entrants jumping on the hybrid cloud backup wagon, it’s increasingly difficult to understand and compare all the options. One critical performance metric that few buyers consider is how quickly data is transferred from primary storage (or from the cloud-connected backup appliance) to the cloud. 

Many customers assume it’s entirely a function of the company’s own Internet connection.  The larger the pipe, the faster the throughput.  While this is largely true, the cloud backup vendor’s own software and technologies can play a large role in improving the speed of cloud replication though WAN acceleration.

As the name suggests, a WAN accelerator is an appliance or software that optimizes bandwidth to improve the end user’s experience on a wide area network.  But, how do they work and what kinds of questions should you be asking of a would-be vendor to get a better sense of how they can boost throughput? 

Here’s what you need to ask:

Do you use traffic shaping or bandwidth limiting?

Traffic shaping technology allows administrators to specify how much bandwidth backup traffic can consume at various hours each day (for example, low bandwidth usage during business hours, high bandwidth usage non-business hours). Traffic shaping also allows vendors to prioritize network traffic based on the type of application or content type (such as email vs. video content). Better vendors will leverage dynamic MTU (maximum transfer unit) re-sizing to optimize throughput by intelligently sizing the amount of data sent on each request depending on your network connection.

How do you dedupe the data?

If you think about it, the most efficient way to accelerate the transfer of information across the WAN is to not send it in the first place. By preventing repetitive information from traversing the WAN, data deduplication can reduce more than 90% of WAN bandwidth. While most vendors will utilize some sort of source-side deduplication to ensure that the appliance keeps only one version of a file, not all cloud vendors will dedupe the data over the WAN (that is, make sure the data being replicated doesn’t already exist within your cloud backup). 

Leading cloud backup companies will use some form of over-the-WAN block-level deduplication.  Technologies like Infrascale’s DDFS-assisted replication query your cloud repository to see if a particular block already exists within the cloud backup.  If it exists, then the block is not replicated to the cloud.  If the block does not exist, then it will copy the data to the cloud.

How do you recover from network disruptions?

Sending data to the cloud is still fraught with network hiccups and packet drops. So, it’s critical that your vendor has the technology to ensure that all of your data gets replicated to the cloud. Leading cloud backup companies, which are steeped in traffic optimization, take advantage of resilient-resumption technology that enables them to quickly recover and continue operating even when there has been packet loss or some other network disruption. Some even deploy more advanced optimization techniques to reduce network latency.

There’s clearly a lot of technology under the hood when it comes to fast cloud replication. 

For backup companies with an appliance lineage, the absence of baked-in WAN acceleration suggests cloud replication is an afterthought. But, by asking some of these pointed questions about WAN acceleration and network optimization, you will be a much smarter consumer of cloud backup services and in a much better position to separate the contenders from the pretenders. 

Dean Nicolls is VP of Demand Generation, Infrascale. Guest blogs such as this one are published monthly and are part of MSPmentor’s annual platinum sponsorship.