Reset on the panacea of cloud

Michael O'Neil, principal analyst, InsightaaS
Michael O'Neil, principal analyst, InsightaaS

Finding the path from initial pain to cloud success isn't a straightforward process, Michael O'Neil, principal at InsightaaS and author of The Death of Core Competency: A management guide to cloud computing and the zero-friction future, told an IT audience hosted by CenturyLink at the St. Andrew’s Club. Yet according to a recent survey by analyst firm Techaisle, fully 96 percent of small and medium businesses listed cloud computing as one of their top priorities for 2015. And a lot of them – 85 percent of business under 100 employees and 94 percent of those with between 100 and 999 employees – already use, or plan to use, cloud computing technologies, trumping even mobility on their corporate to-do lists.

The survey showed that smaller businesses are handling their cloud evaluation differently from the midmarket. While, says O'Neil, smaller companies are fairly equally divided among the three phases of information gathering, identifying potential solutions, and evaluating suppliers, the 80 percent of midmarket respondents said they were in the information gathering phase, with 13 percent identifying solutions and only 7 percent at the supplier phase. He said that the midmarket's approach to the cloud suggests that it already has solutions, and is figuring out how to assemble them into an infrastructure.

Getting there may not be half the fun, but O'Neil offered a twelve step path to guide companies to cloud success.

Step 1, he said, is Pain. No surprise – before you can solve a problem, you have to admit you have one. And many of the top ten pain points for both business and IT can be solved in the cloud.

On the IT side, the Techaisle survey found that budget constraints were the number one issue. Many of the rest of the top ten are caused in part by lack of resources, including maintaining current infrastructure, data protection in its multiple forms, excessive data growth, and controlling costs and improving financial justification for continued IT investment. The business side of the shop's biggest pain points were the need to increase profitability and the desire to grow the business; pretty much everything else in business's top ten revolved around those two issues.

Step 2 is Planning. Compared to the adoption process cycle for traditional infrastructure, some components are a lot less difficult when adopting cloud technologies. While strategic and operational planning both must be done (and are equally difficult for both technology approaches), finding, selecting, implementing, and tracking and supporting cloud solutions can be much easier. Mind you, there's a wide variance in the complexity of the tracking and support; if there's only one service involved, it can be relatively straightforward. However, ongoing course corrections and multiple solutions can add to the complexity.

Step 3 is Requirements (Governance). There is, O'Neil told the audience, a strong correlation between project success and who is actually running it. When IT or the business alone calls the shots, he claimed, the project is likely to go "really off the rails". Collaboration will improve matters.

Step 4 is Supplier Selection. Research in this area showed that investment is a progression, not an abrupt switch. O'Neil said that for every dollar spent on on-premise IT, last year companies spent 11 cents. Over the next few years, he expects the amount to creep up a few pennies at a time, then taking larger jumps starting in 2017 until it hits 95 cents in 2020.

Moving on to Step 5, Recommendations, O'Neil pointed out that small and midmarket businesses have different decision-makers. For infrastructure purchases, he said, in small organizations the business side has the most authority in decision-making, while in the midmarket, IT is more likely to tip the scales. When purchasing new business applications, however, the business side's authority is stronger for both groups.

In larger firms, he explained, there is more separation of responsibility, and IT infrastructure tends to be more complex, so IT influences decisions more heavily. Small organizations may outsource IT management, with decision-making centred in the business.

Decision-makers in business and IT have different priorities, he added. Where business leaders think about agility and data, IT decision-makers consider cost, security, and how to augment existing infrastructure.

Step 6 brings the initial cloud, and O'Neil's key message was, don't stop there. If you implement one cloud, and then stop, you lose momentum.

Based on the research presented at the event, people are taking that to heart. Small businesses reported using five software-as-a-service (SaaS) applications, and are planning to add seven more. Midmarket businesses currently use ten, and plan eight more.

The next six steps are all about growing and optimizing cloud use. For the cloud channel, O'Neil said, data integration between cloud and on-premise applications is mandatory. Without it, the providers will not be profitable.

Orchestration of functions comes as cloud implementations mature. The first step is automation of individual tasks. Next, process orchestration and automation of related tasks generate faster response, better consistency, and better return on IT and staff time. Then, orchestration grows, first across related activities, then into automating entire functions, and finally across the enterprise, with delivery of business infrastructure via the cloud.

Step 12 is the post-cloud era, in which the cloud is not an IT solution, but rather just a part of the business infrastructure. O'Neil thinks that security will be a key driver of cloud adoption, and that the winners will be those who are best at making the transition.

Richard Seroter, head of product management, CenturyLink Cloud
Richard Seroter, head of product management, CenturyLink Cloud

Richard Seroter, director of product management for CenturyLink Cloud, picked up the ball next, discussing how to achieve that successful transition. His conclusion? The model has to change.

Only 27 percent of business leaders were extremely satisfied with their cloud migrations.

First, he said, you have to define your purpose. And without pain, there's no purpose.

"If saving money is the only purpose," he said, "cancel the project." If you're spending more, but it helps launch projects faster and grow the business, it's a win.

"Headcount reduction is also a lousy reason," Seroter said. Cloud lets companies scale without growing the workforce, but it's not outsourcing.

Another bad reason for moving to the cloud is peer pressure. Just because everyone else is doing it, you don't have to (as Mom would say). "If you're not doing it for agility, you shouldn't do it," he said. "You want to look at the environment differently."

The service provider space is very noisy, so you have to look at the real cost of the cloud. Include the cost of migration, training, and administration, not just the cost of the server. There are a lot of hidden costs. And when selecting a provider, Seroter said, "If you don't go hands-on, shame on you!"

He advised the audience to think about what performance means to you. Think about what other services you need. See if you can single source those services. Look at the necessary management toolset – if you choose a service without thinking about ops, you're doing yourself a huge disservice.

Ask yourself if you expect to plug in to third party or internal tools, and if so, check out the API. "If you don't offer an API, don't call yourself a cloud," Seroter said. "The cloud shouldn't be the weird uncle living in the attic; it should be sitting at the dinner table."

Still, support is a concern, and you have to understand your options. Know your SLAs, he advised, and bookmark the vendor's support pages. Know what's included in your fees, and what generates an extra charge.

Allow users to self-serve, and embrace chargebacks; give up control to get control. "Price transparency is empowering," he said. "There is no 'business vs IT'."

His final suggestions were around operations. "Think about making teams partner," he said. "The idea of specialists goes away." Send people to training and conferences to "get them out of their echo chamber". And unlike on-premise installations, where people tend to configure for peak load, he suggested the least possible permanent space, with the ability to scale up and down as required. Applications, too, need to be different. They need to tolerate failure gracefully.

Finally, he said "Embrace automation. It's how you can work in bulk. The whole point of the cloud is agility."