Sometimes the biggest challenge with IT isn’t amassing a lot of capacity. It’s finding a way to most effectively wield it.
By Michael Vizard
For Revlon CIO David Giambruno, effectively managing IT means finding a way to project the power of the data center to the entire company in a way that makes it easier to consume IT resources while also reducing the total cost of IT.
In order to accomplish that goal, Giambruno says keeping things simple is critical.
“We want to make things simple,” he says. “Simple means speed.”
The traditional challenge with projecting the power of IT has been the need to deploy IT resources in every branch office of the company. While generally effective, it’s an expensive proposition. Not only do additional software licenses and IT infrastructure need to be procured, but IT people will eventually have to travel to the branch site to manage it.
Looking to change the fundamental IT math around that equation, Giambruno enlisted the aid of Riverbed Technology to simplify the traditional branch office deployment model. Instead of deploying all those applications and servers locally, Revlon is making use of a Riverbed Granite platform for remotely accessing data.
Riverbed Granite is an appliance in the data center that pre-fetches data at the block level and transfers that data, when requested, to an appliance at the branch office that caches all the data that office needs locally. Riverbed data deduplication technology is applied between the appliances to minimize the amount of data that needs to be sent back and forth across the data center.
“Our goal is to project the data center out to the business,” says Giambruno. “We want to eliminate latency and consolidate data.”
In any type of cloud computing environment, Giambruno notes that latency is the primary enemy because everything stored in the cloud is a file. Latency, he says, “is a cloud killer.”
By caching data locally, Riverbed Granite enables Revlon to transcend network latency issues while centralizing data management in a way that provides multiple business benefits. For example, when a fire damaged a Revlon data center in Venezuela, IT services were delivered via the company’s data center in New Jersey in a matter of minutes. Similarly, when a flood wiped out an office in London, the New Jersey data center delivered application access within minutes. And when Hurricane Sandy devastated the company’s New Jersey and New York data centers, a data center in North Carolina was pressed into service in about 20 minutes.
“It actually took longer to find our SAN guy, who was vacationing on the Outer Banks in North Carolina,” than it took to get the North Carolina data center moved into action, says Giambruno.
Just as importantly, Revlon has reduced the amount of time it takes to deploy an application to about six hours, which represents a 70 percent reduction, says Giambruno. Servers, meanwhile, can be deployed in about 10 minutes and the mean time to deliver a patch update is about 17 minutes, says Giambruno.
Other benefits to this approach, notes Rick Villars, an industry analyst with International Data Corp., is that besides making IT more efficient, it makes it easier to secure data assets in an era of globalization that requires organizations to do business in countries with widely divergent views of intellectual property rights.
“IT needs to make sure data is not misused at the edge of the enterprise,” says Villars.
In the future, Giambruno says Granite will significantly reduce the movement of virtual machines in the Revlon cloud environment. Right now, Revlon is coping with 15,000 automated application moves a month. Going forward, much of the need to do that should be sharply reduced.
“Moving virtual machines adds complexity,” says Giambruno. “I would rather not have to do that because the opportunity for error becomes a lot higher.”
All this simplicity, says Giambruno, is aimed at achieving one strategic goal above all others. “One of the things I live by,” says he, “is to take infrastructure out of the way of doing business.”
This article was originally published on 04-19-2013