Should you really build your own cloud?

In a world where public cloud dominates the headlines, should you bother building your own cloud offering? And if so, why?

CRN spoke to people in the channel who have built their own clouds to get their experienced perspective.

Yes, but with care

The consensus from those CRN spoke to was that yes, building your own cloud can make sense, but you have to be very clear on what you're trying to achieve.

"It provides you with your own product and point of differentiation in a crowded marketplace," said Matthew Moncrieff, managing director at Moncrieff Technology Solutions. "You also have control over the solutions and services that you can bring to market and can adapt them your clients requirements."

"It’s a significant investment," Viatek national sales director James Veness said. "You need to ask yourself 'Do my customers have any specific needs that public cloud doesn't cover?' If you can find where those gaps are, you have a unique selling proposition."

And this is the key point: you can't simply build a cloud offering that attempts to out-AWS Amazon AWS. There is no possible way for an MSP, starting today, to build a compelling offering that is a straight copy of what AWS (or Google Cloud, or Microsoft Azure, or Rackspace, DigitalOcean, IBM Cloud, Oracle Cloud, etc., etc.) are doing.

If you are thinking about building a cloud, you need to have a detailed, nuanced understanding of what you're trying to achieve, and why customers would want to buy from you.

"It's less about technology and more about commercialising an offering in a way that makes sense for your business and your target clients," Sundata managing director Kon Kakanis said. Your cloud offer has to provide value that customers can't get elsewhere.

"Understand why a client would come to you rather than a global player," he said. "Building a cloud includes contracts, delivery mechanisms and demonstrating accountability—on top of building a reliable environment."

There's a lot more to building a compelling cloud offer than simply throwing some technology together, and that takes careful business planning.

"Do your research and if you do go down the path make sure you invest the time to get it right from the start," said Ben Town, managing director at Hosted Network.

"At the end of the day you'll likely be running your customer's entire IT infrastructure and that is a lot of risk and responsibility, not to mention a 24/7 job."

What do MSP cloud customers look for?

If you're thinking of choosing a private, MSP-run cloud instead of a major public cloud, what should you be looking for?

"Larger providers can be more restrictive and potentially more expensive," Town said. "Whereas a smaller provider might be more agile and able to accommodate different setups." If you can't get what you want from the 'one-size-fits-all' approach of a major vendor, the willingness to adjust their offer to suit you might appeal.

"Customers should be looking for differentiated services that meet their specific needs," Veness said. "They should be looking for expertise and high-level certifications and partnerships with key technology vendors, both in cloud and traditional IT areas."

The big cloud providers tend not to be interested in traditional IT, so unless you're all-in with a particular cloud vendor, you'll still need someone to help you with all of your existing heritage IT. Having the same vendor handle both cloud and traditional IT for you could be beneficial.

Sundata's Kakanis also cautions customers to look consider what will happen if they—or their chosen cloud provider—changes their mind. "They should also understand what ‘getting out’ looks like should the relationship falter," he said. This goes not just for small MSPs, but also when the behemoths change direction and discontinue an offering you've come to depend on.

Getting out of a situation that's gone bad can quickly eat up all the cost savings from a move to a slightly cheaper offer, and add a bunch of stress in the process.

The question of which cloud to use is similar to that of deciding whether to build one: Is it really worth it in the end?