Cloud World Series Blog

Navigation Area

Transforming the digital enterprise through Cloud, IoT and DevOps

Cloud World Series Blog


2017: the year that the #Private / #Hybrid #Cloud bubble finally burst and more such predictions Article

Soon the battle may well be between generalist and specialist public clouds.

Bill MewBill Mew, Cloud Strategist, UKCloud (@BillMew)

Many pundits will be offering safe, but obvious cloud market predictions. I wanted to offer a few that challenge current thinking in much of the market: 1) the private cloud bubble will finally burst, 2) without private cloud, hybrid cloud strategies will be brought into question, 3) public cloud will become the norm, but 4) the big players won’t get it all their own way and finally 5) specialist public cloud services will emerge to challenge the global generalists. I may be a contrarian, but I believe that we will see all of this come true in 2017. Here’s why…

It is incredible to see how widely cloud is now accepted as the future of IT, while at the same time there is so little agreement on what kind of cloud is best. Whereas once the big question was whether to move to the cloud, the question now appears to be how to move to the cloud and what type of cloud to move to – private, hybrid, community or public cloud?

I have long argued that the ultimate destination is public cloud and that private and hybrid are simply transitional technologies promoted largely by vendors that are scrambling to catch up – see my earlier musings here – “What exactly is hybrid cloud: A stepping stone, a sanctuary for regulated industries or a rampart for old tech vendors?”

It is refreshing to find that there are other leading experts (far smarter people than me) that are speaking out on this topic as well. UBS Managing Director Steven Milunovich, a notable luminary, recently commented that there is a general consensus that: “private cloud implementations generally are not working, and many companies that begin on a private cloud path end up going down a public cloud path.” In addition, one of the top cloud pundits, David Linthicum, also called this issue out, saying: “Wake up! The private cloud fantasy is over.” Bernard Golden also wrote that The long, slow death of private cloud continues, saying that “At some point the economic unviability of private clouds will become clear. Math will win out.” Furthermore, Subbu Allamaraju, VP Cloud at Expedia, recently wrote a blog entitled simply: “Don’t Build Private Clouds.”

So where did all the private cloud hype come from?

Some question the motivation of the traditional vendors that are pushing private and hybrid cloud. Indeed in their efforts to catch up on the cloud front, some of the traditional vendors attempted to co-opt OpenStack for their own ends. Mark Shuttleworth, Founder of Canonical, recently claimed that the motivation for many was to create a private cloud framework where they could sell their legacy goods. “That bubble is now bursting,” he claims. “We’ve seen a number of retractions from the OpenStack market for whom that strategy didn’t work – it was never going to work but there was a sort of bubble around OpenStack. The smart ones are now coming to the market with genuinely cloud native solutions.” [NB: Here at UKCloud we use Red Hat’s version of OpenStack to provide our clients with a Cloud Native Platform on our public cloud]

If private cloud is a fantasy, then so is hybrid cloud!

Personally I find that one of the most frequent indicators of a bubble is when people start telling you that the economics are different this time – just what we were told with the internet bubble and the debt/property bubble that followed it. The fundamental rules of economics rarely change. Indeed 451 Group recently made just this point. Their research sought to provide transparency across the range of complex cloud pricing models to take into consideration the major factors impacting total cost of ownership (TCO), including salaries and workload requirements. They found that OpenStack and commercial private clouds can compete with and even beat public cloud on cost – but only at scale.

Unfortunately, very few private clouds achieve adequate scale. Even companies of the size of BP are turning their backs on hybrid cloud. BP found that private cloud looked, and smelled, too much like on-premise and it has chosen to go all in on public cloud.

Where does this leave us? Are the global generalists going to get it all their own way? I think not.

Again economics offers the answer. As 451 outlined, economies of scale mean that you need to achieve scale in order to be cost competitive, but this only applies up until the point that scale is achieved. Beyond this tipping point additional scale provides diminishing returns. Eventually the increasing cost of complexity faced by all large organisations comes into play.

In any market, be it retail or automotive or whatever there tends to be a split between generalists that are larger in size and cater for general needs and specialists that are smaller and more focused on the specific aspects of the requirements of individual market segments. You see this with firms like Peugeot Citroen that caters for the general market and Jaguar Landrover that caters for more specialist needs, or with retail outlets like Sports Direct selling general sports goods versus specialists like Snow & Rock that focus on winter sports alone.

bill mew 0301 2

Diagrams: The Economics of Cloud

This is what we are now seeing in public cloud. There are a number of public clouds, such as those managed by significant local players in each market that have achieved scale. They operate at the sweet spot where they are large enough to compete, but focused enough to offer additional value. Some refer to such clouds as community clouds (normally where they focus on a particular local market or niche). At UKCloud, we prefer to categorise ourselves as a specialist public cloud (in our case for the UK public sector) as opposed to the generalist public clouds on offer from the global giants. The main differences between these two types of public cloud are:

  • Focus: In achieving scale both groups benefit from the resulting economies, as well as the breadth of skills to cope either with any kind of problem or indeed with the pace of innovation. The difference is that the specialists are big enough to cope, but still small enough to care, while with the generalists you are one of so many clients that you tend to be just a number. Indeed, catering for a particular niche allows the provider to offer services that are of particular value to their clients and that more directly meet their particular requirements allowing them to innovate or meet their challenges cheaper, faster and better.
  • Openness: the global generalists have each chosen to adopt a proprietary technology stack that is unique to their own infrastructure, while the specialists have mostly opted for technologies such as VMware for legacy workloads and OpenStack for cloud native ones. The global generalists may offer enticing deals to attract you onto their platform, but cost and contract length are not the only things you need to be wary of. See a more detailed outline of the pitfalls in this blog from a colleague of mine: being open about lock in.
  • Sovereignty: In Bernard Golden’s article that I referred to earlier, he claims that in the long run most IT organizations will discover the economic unviability of private clouds, except in a few particular circumstances (e.g. data sovereignty). I agree with him about the viability of private clouds, but would suggest that specialist public clouds with enhanced security and complete sovereignty are the most viable option here. And while Microsoft may well be claiming that they have slain the beast of sovereignty, its recent win in the Dublin privacy court case is currently being appealed and there are many other US regulations that the global generalists, as US companies, have to abide by. They will never get away from the fact that they will always be subject to US law (and the NSA). Local players on the other hand tend to operate from purely local data centres, are subject only to local laws and pay their taxes in their country of operation – tax avoidance being an increasingly contentious issue that we have with the global generalists.

Add to this the currency aspects. With one of the most compelling attractions of cloud being the ability to consume technology as a service and therefore move from capex to opex and thereby a fixed monthly cost, it is kind of self-defeating to expose yourself to currency fluctuations. The recent Brexit referendum caused a drop in the value of the pound which was followed by a 22% sterling price rise from at least one of the global generalists – not something you saw from any of the local providers. Such price hikes are particularly unwelcome if you are locked in, preventing you from moving to a more cost-effective alternative.

So what are the competitive dynamics here:

  1. I can see a compelling argument for a specialist value proposition over a generalist one – especially when the specialist knows and understands my market, while the generalists aren’t even focused on cloud alone (Amazon, Microsoft and Google all have a range of other operations including devices – from laptop and phones to e-readers and games consoles [each with software platforms] – that they are pitching against firms like Apple).
  2. I can see a compelling argument for open standards over proprietary platforms
  3. And I can also see a compelling argument for data sovereignty over data residency (especially after the recent Rule 41 changes in the US).

I would therefore argue that the global generalists won’t be having everything their own way and that the opportunity for specialist public cloud providers is significant. Once the bubble bursts for private and public cloud (which it will), the battle between the generalist public cloud and specialist public cloud players will begin in earnest.

Specialist #public #cloud providers: big enough to cope; small enough to care!


    Share this post:

One Comment

  • Insightful article Bill, thanks for sharing your thoughts.

    • Simon Lydford
    • 3 Jan, 2017 @ 13:38

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>