Getty Images

Save the Children swaps on-premise migration plan for public cloud

International charity Save the Children explains why it abandoned plans to shift core IT systems back on-premise to embrace public cloud computing technology

Save the Children International has shifted most of its back-office systems to the CenturyLink public cloud, as part of the not-for profit organisation's push to serve 15,000 users worldwide.

The organisation focuses on helping children whose lives are disrupted by natural disasters, armed conflict or disease outbreaks, by providing them with medical attention, emotional support and shelter.

Save the Children International has a global staff of 15,000 running humanitarian operations across 55 countries. Workers need consistent access to core IT systems from wherever they are in the world.

This requirement – coupled with the organisation’s never-ending push to conserve costs – largely dictates its IT strategy, Graham Kent, Save the Children’s head of global IT service delivery, tells Computer Weekly.

“We are a $2bn a year organisation, and we need back-office systems like anyone else. We have a HR platform, a collaboration platform, we have a finance ERP system and all those kinds of things,” said Kent.

“The greatest challenges we face are trying to deliver that to our staff – working in a warzone in South Sudan, for example, or a very low low-bandwidth area in West Africa somewhere – with a fraction of the budget a traditional enterprise might have.

“We’re always under pressure to reduce costs on things that aren’t to do with humanitarian programming, because we want to spend our money on saving children, not corporate IT.”

Read more about datacentre migration

Counting costs

Since joining the charity in the summer 2014, Kent has reduced the size of the not-for-profit’s on-premise IT estate by expanding its use of cloud technologies beyond Microsoft’s online business productivity suite, Office 365.

“When I joined, we had a gigantic Office 365 that is very important to us, and we had some accounts with an element of cloud-hosting with Savvis, which has now rolled into CenturyLink Cloud, and we had some of our own server room hosting in London with our own physical kit,” he said.

The cloud-hosting deal was in the throes of being wound down when Kent joined, in response to changes being made to Savvis’s roadmap in the wake of its acquisition by CenturyLink in 2011.  

“As I arrived, I was told that one of the teams I inherited was in the process of migrating services within Savvis, and they were bringing them back to go in-house to host them on-premise again, because they were under pressure to save money, and they deemed they could lose this operational expenditure cost,” Kent said.

After spending some time settling into the job, and getting a feel for the organisation’s IT setup, the Savvis migration plan began to weigh heavily on Kent’s mind.

“I became increasingly concerned this was not a good move. In my last job at a we ran our own stuff on-premise because we had a team of 50 great people doing loads of DevOps and it worked out much better value for money,” he said.

“We have a much smaller team here because people are such a big cost to the business, and I started to look at this project and the migration was not doing well; the guys were trying to achieve a $50,000 a year saving, but I started to worry this move would put us in a terrible position long-term.”

Arrested development

Three months after joining, Kent called a halt to the initiative to carry out a thorough total cost of ownership and performance analysis. Kent wanted to see what the long-term benefits would be for Save the Children International.

“We went back and looked at the project from a ‘bigger picture’ perspective. We came out with a model that I did not want to bring that big back in-house – it was going to cost us more, and would require more staff, and I was extremely concerned about the performance,” he said.

After completing this fact-finding mission, Kent concluded the move would save the charity $50,000 in the short-term, but – as the demands placed on this hardware grew over time – another costly migration was likely to be needed within 18 months or so.

Kent shared his misgivings about the long-term implications of this move with the charity’s senior management team, and put forward an alternative business plan that would eventually pave the way for nearly all of its major IT systems to run out of the CenturyLink public cloud.

“The view of the CFO and the CIO was that we’ve been underinvesting in IT and we want to use IT-enabled change within the organisation over the next three years,” he said.

Sifting public cloud suppliers

“So, I wrote a business case back up to the CFO and CIO – which was difficult, because the budget was already locked in for 2015, and it is very hard for a non-governmental organisation (NGO) to change that kind of thing.”

The business case suggested abandoning the on-premise migration, and focusing instead on shifting the organisation’s human resources (HR), customer relationship management (CRM), donor management and staff intranet to the public cloud.

Before opting for the CenturyLink Cloud, a number of other public cloud providers were considered and discounted from the running over a four-week period because they could not meet the organisation’s support requirements, said Kent.

“I’ve run AWS and Microsoft Azure before, but what marked CenturyLink out for an organisation like ours is the human-side of the account management that comes with some of their services, because you’re not just an email address at the end of a support line, which you are with AWS unless you pay for premium support.”

“Whenever we do purchasing of IT services nowadays, there is a cost importance but service is also important to us because we don’t have big 24/7 teams.”  

Joining the dots

The CenturtyLink contract was signed at the end of 2014, and the phased migration of Save the Children International’s systems was completed in February 2015.

“There is one primary system we haven’t moved, which is still physically located in the same server room that this was all originally going back into. We’ve been working with them this year about migrating that, but it’s a much more complex undertaking,” he said.

“We haven’t moved that one yet because it’s a political decision across the entire association about whether or not we should all merge all the instances we have of this in different countries before we do a migration.”

Moving the bulk of its back-office systems to the public cloud has benefited Save the Children International in a number of ways, and freed the organisation’s small IT department from getting bogged down in time-consuming infrastructure maintenance and support work.

“It has reduced the amount of time my people have been spending on managing the core infrastructure, and I’ve been able to redirect them to look at lots of the legacy infrastructure we have in field offices and country officers around the world that we’ve inherited over the years. They can look at stuff, which is much more strategic to our organisation.

“The other thing that’s been useful, now we’ve won the argument about swapping from the Capex to Opex investment, is that I’m able to much more accurately financially model what I’m going to do which makes me more popular with the finance team.”

Read more on Infrastructure-as-a-Service (IaaS)