An RPO By Any Other Name…

RoseI spent last week in London for IP Expo. During the show I led a couple of sessions, one on big data mobility, and another on replication acceleration. Though I plan to focus on Big Data in another entry, for this post I want to focus on replication and data protection.

One thing I noticed while in London was the difference in terminology we use — not just “routing vs. rooting”, but really how we talk about data protection. When I talk about replication acceleration, the main focus is on meeting your recovery point objective (RPO). In the US this message generally works, and storage administrators think about RPO as a key indicator of the health of their data protection scheme. In London, I realized that “RPO” is not a driving factor – but at the same time, it is.

As I was talking to our London based SE about the replication accelerator messaging, the focus on RPO, and why Americans use “z” instead of “s”, he mentioned that his customers don’t care about RPO. Needless to say I was shocked to hear that RPO is not important!  However, as we continued to talk we both realized it was a mismatch in terminology — in London, customers are focused on meeting their “backup windows”, even when the data is replicated. Thankfully, we had this conversation on the setup day and I had time to edit my slides before the session.

In North America I talk specifically about Recovery Point Objectives, and how difficult it can be to meet them with the deluge of new data. This EMC report helps illustrate the data growth problem we are all facing — data is growing faster than ever before, and we keep more of it for longer periods of time, usually forever. Protecting this data via replication is common for companies that have demanding requirements for data loss and recovery times. Helping customers meet their RPOs is what the Velocity Replication Accelerator (VRX) is all about.  But when a customer doesn’t recognize “RPO” as a requirement, they might miss the value that the VRX software provides.

How do you talk about RPO if someone doesn’t actually have one? You talk about “backup windows” and “replication times”. Typing this out makes it seem very obvious, and it should be, but it is different in conversation — I didn’t realise (that soft s is for my friends in the UK) that RPO didn’t resonate…but the requirement still exists….they just call it something different.  This is what I discovered in London last week — I did a lot translating in my head every time I wanted to say RPO to make sure that I said “replication time” or “backup window”. Communication problem solved.

On to the task at hand: replication acceleration and how it can help.

The objection to replication acceleration most prevalent in Europe is the relatively short distances they have between sites. One customer told me that they replicate all the way across the country…..all 200 miles of it.  And because the distance is so short, there isn’t a latency problem, so what is there to accelerate? The answer comes down to quality. If you can afford a private network connection with no loss or congestion, and you are going a relatively short distance, you don’t need an accelerator. There it is!  A product manager just told you that you don’t need his product.  BUT, yes that is a big “but”, if you don’t have a private line, or you have congestion or loss…you do have a problem.  Network quality slows replication and backup traffic across the network and will cause jobs to exceed the time allotted (RPO). This happens even with low latency. You can read more here about a customer with low latency that reduced their replication time from 10 days to 5.5 hours.

There is also the use case of the customer that has a private line and is trying to save money, so they want to replicate over the Internet. Five years ago this would be very difficult, and probably unsupported. Today, with Silver Peak’s VRX software, this is not only possible, it’s easy.

All in all, once we got the terminology problem fixed, and addressed the limited distance concerns, we discovered that there were a lot of attendees who could use a solution to fix their backup and replication problems across the network. For those individuals, I said the same thing I always say: “go to and download the software. Try it on your network, with your data, and your backup or replication window.  And oh yeah, if you do try it we will give you full support, 24/7, for 30 days.”

This is the last post of tradeshow adventures until next month when Silver Peak attends NetApp Insight EMEA. Stay tuned for a summary of my session on big data mobility, and why you should move big data even though your storage vendor might not agree.  Then it’s going to be all about deduplication and compression…and why you are wrong when you think that you don’t need it on the network when your storage already does it.

Cheers, E

(and Tottenham Hotspur are going to the Premiere League trophy!)

Image credit: WikiMedia Commons