A substandard data centre that leaked water and had air conditioning failures forced NHP Electrical Engineering Products to move its entire IT infrastructure to the cloud.
Vic Wotherspoon, general manager, information technology at NHP discussed the organisation’s journey to infrastructure-as-a-service over 18 months during a panel session at last month’s CIO Summit in Sydney.
NHP’s headquarters in Richmond, Victoria – which housed its data centre – suffered rain leaks in the winter and air conditioning failures in the summer.
“It was getting to the point where we felt we had to move our equipment elsewhere. We simultaneously had just completed a process of standardising our IT infrastructure and making sure the software was integrated and virtualised,” Wotherspoon said.
Having that integration done first before moving all of its production systems to a Logicalis cloud under a three-year contract was crucial, he said.
“If we had deployed software-as-a-service for each of the individual point solutions that we had in our organisation and then subsequently tried to integrate them, we would have [had] all sorts of problems,” he said.
“So having that integration done up front and moving everything over was the best approach for us.”
The move has enabled NHP to cut costs by moving from a capex to an opex model for its IT infrastructure, and roll out disaster recovery as a service.
It’s OK to miss dates
During the initial stages of the project, NHP gradually moved its non-critical systems across to the cloud.
“We did notice along the way that we had a problem with latency … with the exception of that and we missed a couple of dates, the project went really well,” he said.
Missing implementation dates for this initiative was not an issue, says Wotherspoon, who advises organisation not to be a slave to deadlines if you don’t have to be.
“I was quite happy to miss a couple of dates along the way because we were just not ready,” he said.
Wotherspoon admitted that the organisation should have done a lot more rigorous testing on response times.
"My team were fairly confident that latency would not be an issue when in fact it was. That was one of the reason why the date moved. It was a bit disappointing but it wasn't the end of the world because in this particular project if it didn't work, we'd just move it back – but we actually didn't have to do that.
"Because we moved the virtual servers over a number of weekends, we only started to see the latency problem developing gradually. While it was a little disappointing, it wasn't that big a deal, but if we had put a little bit more time in upfront, that would have been the main lesson learned," he said.
Daniel Sultana, CIO of Logicalis – who was also on the panel – added that it would be naive to think that a project this size would be completed without any issues.
"The latency issue was resolved within a couple of weeks and it was multi-faceted from application through to the network. The mitigations were in place at that point in time to say 'can we fix this, yes or no? What's the impact on the business?'" he said.
Cultural implications of cloud
Moving core IT services to the cloud did create a degree of anxiety among NHP’s technical staff, said Wotherspoon.
It’s important to explain to technical people that it doesn’t really matter where servers are located, their job function is the same, he said.
“The main reason we were able to overcome that [anxiety] was that our technical people and the technical people supplied by Logicalis on the project were operating as one.
“It’s important to educate the technical staff and explain to them that you are not trying to remove them from the organisation, you’re just trying to move your boxes elsewhere.”
Logicalis' Sultana said there's a substantial culture shift that infrastructure managers need to get over.
"Infrastructure is not an asset, it's a liability. The data is the asset so if I take Vic's previous environment, that was a liability. Waiting for it to get flooded or catch on fire; we've got to get past this concept that 'I need to see flashing lights in front of me in my data centre or else the business can't function,'" he said.
"So, I think we have that leap in maturity we have to make if we are going to get to the next level."
Managing contractual requirements
Wotherspoon said NHP has a checklist for all contractual issues with IT suppliers but making sure terms and conditions in cloud contracts are met comes back to relationships.
“In some cases, depending on the supplier, going through the terms and conditions and the contractual issues can be quite cumbersome,” he said.
“Because of the long-term relationship [with Logicalis], it was a bit more straightforward. We already had NDAs in place and a number of successful projects that we had completed with Logicalis.
“So for us it came down to what the individual deliverables were for the contract and the financial aspects of the payment points throughout.”
Follow Byron Connolly on Twitter:@ByronConnolly