Statistics Canada makes room for lots more numbers

If data were girth, Statistics Canada would be expanding at an alarming rate.

At an average growth of five per cent a month, it was quickly getting to be too big for its storage britches. Something had to be done, so it turned to Symantec‘s Data Centre Foundation solution, a mix of data and storage management tools, to help it consolidate its environment.

StatsCan, which has gone from 160 servers to two NetApps appliances, is a mere shadow of its former self, infrastructure-wise.

Since StatsCan began the project in the summer of 2004, its storage infrastructure has grown by 842 per cent, a combination of consolidation and normal growth of its data assets, said Guy Charron, assistant director of infrastructure solutions at the federal department.

His organization, which he describes as a mini Shared Services bureau, is constantly on the hunt for ways to increase efficiency and lower costs to clients. The organization can’t make money – but it can’t lose any either, he said.

“We propose our rates at various committees at StatsCan,” he said. “We’re under a microscope with regards to our costs around the delivery of IT services to our clients, so we are continually trying to find ways to lower costs to clients.”

Housing all its data on Tier 1 disk arrays, which is what it had been doing, was not one of those ways, but information lifecycle management looked promising, Charron said.

“Of course the product we were looking for had to have some integration to our backup solution, which at the time was Veritas,” he said. “And having decided to go with NetApps appliances, we needed products that linked to those technologies as well.”

NetApps has partnered with Symantec on enterprise storage offerings, making that integration easier, he said. 

The project is now moving out of the labs and into deployment by the first quarter of next year.

“We’ll roll it out to informatics branch first — the infrastructure group and our sister division on the development side,” Charron said. “We’re testing the functionality and at the same time because we’re using a mirrored copy of our production environment we’re fairly confident of our findings and the efficiencies we’re going to achieve.” 

Early indications are impressive: according to a Symantec business value assessment, the cost per gigabyte of data has been reduced to $1.50 from $5.74, for a total savings of $10.83 million. There are fewer IT employees manning the backup operations – 15 as compared to 28.5, with no plans for any increases through 2008, and a labour savings of an estimated $392,000.

Relocating data is expected to save $791,000 from 2006 to 2008, and the improvement in storage utilization is projected to increase to 90 per cent from 51 per cent for an estimated $6.5 million cost reduction between now and 2008, and an additional $16.9 million in 2009 and 2010. 

The next – and most challenging — part of the project will be getting users to profile their data – setting rules around how long they want to keep it before deletion, how many copies of files they want and whether or not the data should be archived, Charron said.

“We have a requirement at StatsCan to keep our Census data for 92 years, but we would like our clients to at least apply some kind of guidelines to how long they want to keep their other data as well,” he said. “The technology is the simple part.” 

To help users with those decisions, the department has set up a new committee to look at the rules related to mission-critical applications and data, he said.

Craig Andrews, director, system engineering at Symantec Canada, said the difference between Symantec’s toolset and those from other vendors is that the data centre foundation provides a set of tools that works across all operating systems.

Not having that interoperability is like having to hire a different mechanic for every make of car in a car repair shop – it’s not very efficient, he said.

“The big advantage to the company using our tools is a reduced cost of operating their data centre because they don’t have to have people trained on multiple tool sets that do the same thing.”


Share on LinkedIn Share with Google+