Hoping to take some of the stress off its groaning “big iron,” Canadian Pacific Railway will soon be rolling out a hybrid cloud.
CP realized the company’s legacy data centres couldn’t keep up with customers’ demand for self-service applications, including those accessed by mobile devices, so it began to build a private cloud for its Canadian operations, which is linked to the Amazon public cloud, serving its offshore interests.
“One of the biggest impediments we’ve had to acting in IT is the lead times to get new infrastructure for development, for test, for experimentation purposes as well as production purposes,” said Stuart Charlton, executive IT advisor at CP.
The company has several data centres, running a mix of mainframes and mid-range systems. As a massive, long-established company constrained by the weight of its on-premise backend infrastructure, CP’s IT department often had to “slip things underneath it” when new projects came up, Charlton said.
He describes the traditional way IT is consumed as a “tax,” or demand driving supply. Invariably, he says, IT will constantly have to react to increased demand.
“With a private cloud you’re separating the supply from the demand,” he said. “You’re purchasing capacity independent of the demand and then you’re tracking the trend.
“So, it’s a separate process—it’s different. The demand no longer drives the supply. You’re just predicting it and you’re buying the capacity ahead of time because you’re making the assumption that it’s cheaper to have excess capacity on the floor just in time than it is to wait. Because waiting is the killer.”
“We really knocked it out of the park,” he said.
Charlton added that CP is using IBM WebSphere eXtreme Scale “heavily” for its software development in the cloud. eXtreme Scale was developed for distributed caching in cloud environments.
The railway’s public cloud is being used by several hundred developers in Calgary and overseas, in India. Its data centre in Singapore serving offshore customers has a secure link back to headquarters in Canada, Charlton said.
Jorge Garcia, an analyst at Montreal-based Technology Evaluation Centers Inc., said the way the railway is approaching cloud computing is forward-thinking compared to the way other large enterprises are doing so.
“As far as I have seen, it’s more like a reactive process, not a predictive process,” he said.