Restaurant fills DR gap with all-flash Pure and dodges higher cloud cost | Computer Weekly

US restaurant chain Red Lobster opted to standardise on Pure Storage flash storage arrays as it tried to fill a gap in disaster recovery (DR) provision that would have left it severely exposed in case of an unplanned outage at its primary datacentre.

The move to fully on-premise mirrored datacentres also meant it dodged cloud operational expenditure (opex) costs that would have been six times the cost it has incurred on-site.

Red Lobster specialises in seafood, in more than 500 restaurants across the US and Canada, as well as take-out and delivery services.

It runs its IT from two Equinix datacentres, one in Ashburn, Virginia, the other in Dallas, Texas. Core IT revolves around point of sale, internal applications such as HR and finance, and procurement, all of which underpin the company’s primary operations – buying ingredients, hiring staff to cook and serve food and drinks, and taking payment.

When CIO Shawn Harrs came to Red Lobster, the most urgent task was to build out a gap in DR provision that meant the company’s primary datacentre did not replicate to the other.

In the process, the company also end-of-lifed storage infrastructure from four incumbent suppliers and standardised on Pure.

“When I joined the company two years ago, the primary datacentre in Ashburn did not have failover capability to Dallas. It had never been built out,” said Harrs.

“When I came into the company, I did an assessment across all the capabilities of the company,” said Harrs. “And as I dug into it, I realised the company had built a pretty strong infrastructure capability in Equinix datacentres, but it never completed the project. It stopped well short of providing a true business continuity capability.”

Harrs built his business case and presented it to the Red Lobster board.

“I needed to implement a capability where if everything running and live and being used fails, there has to be an equivalent exact copy of it readily available to run. There were different ways to accomplish that.”

When finance looks at the lost revenue of one day compared to the one-time cost … and a one-hour return to service, it is a very easy sell

The options evaluated initially were all cloud and on-premise, which included the use of spinning disk. But the initial iteration of the latter option did not give a quick enough restore time, said Harrs.

“We looked at a cloud option, and we looked at an on-prem option that didn’t return the business to operational functionality within a time frame that was acceptable,” said Harrs. “Another option we looked at – which I called the Goldilocks option – was to supplement that second option with Pure Storage because of the speed at which we can replicate what’s happening in one datacentre to the next.

“From a cost perspective, the options were highest cost, lowest cost, and then [Pure] in the middle,” said Harrs. “It’s sort of the middle cost, but it gets us back up and running in an hour versus in a day. And so, when finance looks at the lost revenue of one day compared to the one-time cost, or one-time and the ongoing cost of a one-hour return to service, it is a very easy sell.”

So, Harrs’s team deployed Pure Storage FlashArray storage at both sites, with replication between them to allow for failover. Capacities run to petabyte scale.

Red Lobster opted for Pure’s Evergreen One storage as a service, which allowed the company to spread costs.

“I wanted to get a favourable mix of one-time capex [capital expenditure] and recurring opex investment,” said Harrs. “To invest a significant amount of capex in a one-time initiative and depreciate it over a five-year period would have made the project very heavy in that one fiscal year. With this option, I was able to take the total cost of the project and spread it out over a three-year period.”

Has Harrs been able to quantify the benefits of moving to Pure Storage infrastructure?

“Yes, in sleep. Because I can sleep at night. It’s the kind of investment that you never want to know about. If you’re a CEO or a CFO or an operator, you want to know that you’re going to be able to open your restaurant tomorrow and it’s going to be able to run. This can give them that certainty.”