Site icon Bridgeworks

What’s wrong with Data Domain replication?

Andrew Martin, Technology Director of Data Protection and Life Cycle Management company, SymStor shares his insights about Bridgeworks WAN technologies disrupting markets.


October 4, 2018

 

 

What’s wrong with Data Domain replication?

Give me a break, I’ve not been that vocal for a while and to be honest, I have had good reason to be silent, myself and my team have been very busy testing our socks off.

So, before Dell EMC point both barrels at my head and pull the trigger…. HOLD YOUR FIRE…

This is for all Data Domain customers who are using replication, be it NetWorker CCR (Clone Controlled Replication), MTree replication or any other application that is replicating data to another Data Domain between data centres regardless of geographic location. We accelerate Data Domain replication time by around 95% over your WAN link delivering near LAN speed. Why is this important?

It’s not about data availability anymore, it is to do with data recoverability and the speed that you can restore data over your WAN link. I have witnessed first-hand where customers are up to 5 days behind their replication target, I find that quite shocking and suspect that this is unacceptable from a business prospective. I wonder how much it would hurt that business if they suffered an outage?

All WAN links have latency, that’s a fact. Every WAN service level agreement has 0.1% written into the small print, trust me I have checked quite a few from various Service Providers. With our partner Bridgeworks, we will maximise your WAN data throughput regardless of the data. We do not care it is encrypted, compressed or deduplicated. We allow you to accelerate your data across the WAN at up to 98% of the WAN capacity you invest in, month after month.

Let’s get techie for a moment and set the scene for this real-life experience. 1GbE virtual private circuit (VPC), Latency 40ms and that fluctuates as does packet loss. iPerf between both Data Domains achieved 821-906Mbits/s

Looks OK right? Not really, when we looked at the statistics of the actual Data Domain interfaces they were only achieving on egress of 11.1 to 34.39MB/s and ingress between 7.59 to 15.65MB/s on the receiving end. We even spent time tweaking the backup application, including RPS cloning and manipulating the parallel streams on the Data Domain. Compression, data deduplication and encrypted data does not affect how we accelerate your replication traffic.

What happens next, you end up with replication lag with terabytes of data waiting to be sent. This can equate to days of replication lag. Remember, the bottleneck is not your Data Domain’s. It is your WAN and we can fix this for you without breaking the bank. We are able to stand-up a virtual machine as a PoC supporting up to 10Gb within an hour.

Contact me if you wish to have a demonstration, it will make your eyes pop and jaw drop!

Click here to read Andrew’s post on LinkedIn.

 

 

Exit mobile version