It’s a popular deployment – something that all the analysts agree on. According to Flexera’s 2019 State of the Cloud survey, 84 percent of enterprises now have a multi-cloud strategy. A Smarter With Gartner article confirms this; “in a recent Gartner survey of public cloud users, 81% of respondents said they are working with two or more providers.”1
But, at the same time, we’re seeing fewer and fewer cloud operators being chosen. Companies may be choosing two or more providers, but they’re being selected from a shrinking list.
While this is good news for AWS, Google and Microsoft (although AliBaba is beginning to make more of a splash), they don’t all get to have things their own way, because not all big players are equal.
Organisations need to make a judgment on their cloud usage as they all have different attributes, strengths and weaknesses. For example, Google is particularly strong when it comes to data analytics – the company’s BigQuery is extremely flexible and supports a high degree of scalability. And, of course, any organisation dependent on Microsoft may find Azure, with its integration with the company’s existing enterprise software, an alluring prospect.
And that’s without considering other, specialist providers (possibly vertical or sector-specific) that may be deployed with the public cloud providers.
Some have technical reasons, some have legal, but one of the prime factors is geography. Some cloud providers are stronger in particular regions, possibly in terms of support, but also due to latency so where a providers’ data centres are situated could well have a part to play in the decision.
Then there are the charges. It’s not necessarily that one provider is cheaper than another, but that they have different price structures.
Companies may also opt for multi-cloud deployment because it offers efficient and secure disaster recovery. For many, a data outage at a provider could be financially crippling – as for some companies that couldn’t recover after an Amazon outage in April 2011 when a few companies couldn’t recover after several days of disruption – so the added protection of an extra supplier could offer peace of mind.
As for governance, countries have different regulatory requirements and privacy regimes and just because a company has a clean bill of health in one, doesn’t mean it does across all they operate in.
Crucial to any multi-cloud deployment is to make sure your organisation has visibility and control across your cloud environment. This means considering how best to manage the cloud - whether you have the necessary skills in-house, or you need a service provider to manage the complexity.
A development that may help is creating a data fabric across your multi-cloud environment - a single source of data across separate data centres including AWS, Azure or private cloud. This overcomes differences in how data is stored, used, and accessed by different clouds.
One of the key things to bear in mind is that multi-cloud isn’t cheap. It’s often presented as such, suggesting that user organisations can play the various providers against each other and negotiate better deals but, in practice, this rarely works out.
There are several reasons why. Firstly, as we’ve already seen, different providers have different pricing structures, so you can’t readily compare apples with apples. Secondly, not every company has an all-embracing cloud policy: one department may be buying from Google, another from Amazon. Even with Amazon, one department could be using reserved instances, while another doesn’t.
And there’s an even more pressing problem: an expert in AWS isn’t necessarily an expert in Azure. Cloud skills aren’t always transferable. A business going down the multi-cloud route may find itself recruiting many more people than originally planned. And applications that have been provisioned for one cloud might not easily be deployed on other.
But there’s a way around this, one that’s increasingly being deployed: the wholesale use of containers across an organisation can make it easier to work with multi-cloud providers. The whole ethos of this technology means it’s easier to port software across different suppliers.
There’s an overhead though. There aren’t, as yet, huge numbers of Kubernetes architects in the world, so companies adopting this approach will likely pay a premium. And this may mean rethinking an entire IT strategy, as well as carefully considering the options. There are a host of reasons for wanting to go multi-cloud – the flexibility, the security, the optimisation of some applications – but there’s a cost involved. Isn’t there always?
But as long as the CIO has considered all the implications and not rushed into the process, adopting a multi-cloud approach could be very beneficial. Many early cloud providers have fallen by the wayside or been acquired. The availability of a smaller selection of major providers has provided a degree of rationalisation, making it easier for customers to choose.
And it’s this dominance of hyperscale providers that has provided much of the impetus towards a multi-cloud approach, says Michael Warrilow, VP analyst at Gartner. “Most organisations adopt a multi-cloud strategy out of a desire to avoid vendor lock-in or to take advantage of best-of-breed solutions. We expect that most large organisations will continue to wilfully pursue this approach.”1
So the key to making sure your organisation’s multi-cloud strategy serves you well – both now and in the future - is to avoid rushing in. Make sure your plan has a measured approach at its heart – carefully considering the management, pricing structure and skills necessary to make your organisation’s digital transformation a success.
1Gartner, Smarter with Gartner, Why Organizations Choose a Multicloud Strategy, Laurence Goasduff, 7 May 2019 https://www.gartner.com/smarterwithgartner/why-organizations-choose-a-multicloud-strategy/