I hate cheesy marketing slogans as much as the next guy… but when it comes out of a customer’s mouth and they mean it, there’s no better time to pull out their cheesy slogan and put it to use. 🙂
This customer was typical in the sense that they had a lot of databases, a few of which were deemed “mission critical,” and they had no DR plan in place except for some duct tape consisting of backup to disk – images shipped remotely. Ugh.
Their business is online. Lots of OLTP, lots of high random IO (great candidates for EFD), and lots of large (multiTB) SQL databases It was a challenge to get backups completed in their nightly windows, and it was a challenge to find the right solution for long-distance disaster recovery. They wanted continuous replication and their backup-to-disk images shipped over the network wasn’t cutting it. Their network pipe was a T3 (which isn’t very large these days) and they couldn’t get 5TB of data sent to their DR site (2000 miles away) in a reasonable timeframe. They needed to fix it.
The group came up with a stated RTO goal of 24 hours and an RPO goal of 4 hours. And off Aaron (sys admin) went to compare host-based options to array-based options and network-based options.
Aaron said he doesn’t like putting more stuff on the hosts. It only leads to problems. He’d rather dedicate his servers to working for the application and not for backups and DR. How many times does your hosts have issues related to funky drivers, or just not want to start randomly… Not often, but enough to consider keeping the hosts clean.
They have a DMX on the source side and a CX array on the target side. They needed something that wasn’t tied to a particular platform – he wanted heterogeneous replication. So the array based option was thrown out. EMC has other solutions that involve more of a push/pull of a large chunk of data – but that’s not really continuous replication, is it?
To quote Aaron,
“RecoverPoint was only complete solution to meet all our requirements; we were unable to meet our DR objectives until RecoverPoint.”
or this one…
“Without RecoverPoint we wouldn’t have a DR solution in place.”
After the one-business-day installation was performed by an EMC consultant, all has been good. They’ve been getting some decent throughput on the RPA’s (RecoverPoint Appliances) and pushing about 5TB of data from Oregon to Texas (they are tripling this in the next few months). He loved the surprise compression benefits (5-15x depending on the data), which lets them push more data with less bandwidth.
He also said a couple things that really stuck out in my mind related to operational efficiency and cost. For him, it wasn’t strictly about the purchase price but about the overall value that this technology provides. Not only does it provide a simple DR solution for their most critical business data, it also minimizes the amount of time spent working on DR.
I will finish with the full version of the blog post quote that really struck me:
“IT people don’t have a lot of time to work on DR, this is a set it and forget it solution”
You can’t make this stuff up. Gotta love RP.
I’ve put some RecoverPoint Screenshots over on SlideShare – go check em out!
Be sure to click the little Full Screen button, otherwise you won’t be able to see much.
- EMC Recoverpoint-Design Considerations (Bit III) (virtualtacit.com)
- The End of Backups..the CDP talk (virtualtacit.com)