(A data centre in Hamina, Finland. Source: Google)
Not long ago, European customers of the global public cloud vendors relied upon a single data centre ‘region’ for all their cloud computing needs. From Lisbon to Lviv, Kiruna to Kalamata, customers of Amazon Web Services (AWS) and Microsoft Azure sent everything to Ireland, and customers of the Google Cloud Platform (GCP) sent everything to Belgium. And, mostly, public cloud’s early adopters in Europe just got on with it.
For the majority of public cloud workloads, storing and processing data somewhere in the European Economic Area (EEA) really was — and is — good enough. Network latency was mostly low enough not to be a problem, and European regulations covered the main use cases well enough to appease all but the most cautious lawyers.
But connections can always be faster, and there are still use cases in regulated industries and government where keeping personal data inside specific geographic borders is either essential or encouraged. And, more and more often these days, customers just seem to feel happier when their data doesn’t leave the country. Mostly, no law requires it, and no regulation recommends it. But it’s still happening. We should all be pushing back against this odd trend towards data balkanisation, much harder than we are.
The hyperscale global clouds seem to crop up pretty much everywhere, these days. But we all know that customer requirements differ, from industry to industry, and from country to country. So... how do they cope, and how do we account for the peculiarities of different markets?
The report (my first Wave, so allow me to feel pleased with myself) is, of course, interesting and useful in and of itself. But what's more interesting, perhaps, is that it's part of a collaboration that allows Forrester to account for those regional quirks.
How often have you been told you can't use a mainstream public cloud provider? Quite often, probably, especially if you happen to work in a regulated industry like banking or healthcare. And what justifications are you given? The regulator "won't let you," no doubt? That's a good one. And "it's not secure" is often pretty close behind. Either that, or the argument that generic public cloud infrastructure can't possibly meet your very special, very unique, very carefully crafted mix of requirements?
Sadly, despite the frequency with which they're trotted out, these attempts at justification stand a pretty good chance of being either hearsay, or just complete nonsense.
It's easy not to change, and to justify your inertia with reference to the scary, punitive, hopelessly luddite regulator. It's easy to continue lovingly polishing the hideously complex snowflake your internal computing environment has become. It's far harder to look at the truth behind the hearsay, and to work out when doing something different might — or might not — be the better approach for your business, and its effort to win, serve, and retain customers.
(Confusing messages. Image by Wikimedia Commons user 'Melburnian')
Again and again, we hear examples of companies struggling as they try to realise the benefits of moving to cloud. They know what they want to achieve as a business, they know that cloud can help, but they cannot translate that understanding into the way they specify, procure, and run the technology.
There are plenty of organisations willing to help, offering everything from design and migration services through to management of infrastructure and applications on an ongoing basis. Even in the public cloud world, it's easy to find companies eager to take your money, and then start and stop workloads on your behalf.
"Cloud computing changes the way that applications are designed, built, and run. It is often part of a broader organizational change, as enterprises move to embrace digital opportunities. Providers of managed cloud solutions need to recognize this shift: They must do more than simply run a customer’s computers. But CIOs seeking a trusted partner to assume this broader role find that too many managed cloud offerings fail to rise above basic management of infrastructure."
Two weeks on, the result of the UK referendum on membership of the European Union (EU) continues to reverberate around the world. Forrester provided advice for clients needing to understand the business implications. Looking at the specific impact on public cloud deployments in Europe introduces a number of additional points. These are best considered in three separate contexts:
that of companies wishing to serve customers in the UK
that of companies wishing to serve customers in the remaining 27 EU member states (the EU27)
that of companies wishing to serve customers in the EU27 from a base in the UK.
When we think about the public cloud, the list of credible providers can sometimes seem rather short.
(The Great Wall of China. Source: Paul Miller)
In North America, Europe, and elsewhere, the same few names tend to dominate. But not in China. There, big local brands continue to command impressive market share. And now they're looking to expand into new territories, including Europe.
Huawei hardware and Huawei's distribution of the OpenStack open source cloud platform power T-Systems' Open Telekom Cloud. This was launched, with some fanfare, at CeBIT in Hannover.
Alibaba Cloud, which leads the Chinese public cloud market, is also coming to Europe this year.
In my latest report, I take a look at what both Alibaba and Huawei bring to Europe's public cloud market, and ask whether they can repeat their domestic success in this market.
TL;DR - it would be unwise to discount either of them.
First we need to define what we mean by “enterprise cloud”. For this definition, the minimal criteria set includes: robust security, reliable performance, disaster recovery, growing set of services, constantly investing and a great and growing ecosystem of partners. Based on this definition (along with the tremendous growth in public cloud), then all of the public cloud leaders are indeed “Enterprise-class”. In short, the term "Enterprise-class" is fundamentally a term targeted to allay the cloud fears of enterprise technology managers.
Providers are adapting to offer managed services for the megaclouds.
The big public cloud providers, most of which are still from the United States, sometimes have a hard time finding ways to balance their legal obligations at home with the quite different sensitivities they encounter amongst their new international customers. For a long time, the toolkit has been pretty consistent: site data centres as close to the customer as possible, vehemently support political efforts to harmonize laws, and ocassionally be seen to stand up to the worst execesses of Government over-reach.
(Source: Flickr user Luigi Rosa. Image licensed under Creative Commons Attribution License)
Microsoft's announcements in Germany today appear, on the surface, to follow that model pretty closely. But there's a twist that's potentially very important as we move forward.
First, the standard bit. Microsoft, yesterday, announced new data centres will be operational in the UK next year, joining existing European facilities in Dublin and Amsterdam. Big competitor Amazon did much the same last week, announcing that a new UK data centre will be online in the UK by "2016 or 2017." Given the vague timescales, it might be easy to assume that Amazon was trying to steal a little of Microsoft's thunder with a half-baked pre-announcement. And then, today, Microsoft announced two new data centres in Germany. Amazon already has a facility there, of course.
I was in Tokyo last week, for the latest OpenStack Summit. Over 5,000 people joined me from around the world, to discuss this open source cloud project's latest - Liberty - release, to lay the groundwork for next year's Mitaka release, and to highlight stories of successful adoption.
Tokyo's Hamarikyu Gardens combine old with new (Source: Paul Miller)
And, unlike many events, this wasn't a hermetically sealed bubble of blandly anodyne mid-Atlantic content, served up to the same globe-trotting audience in characterless rooms that could so easily have been in London, Frankfurt, or Chicago. Instead, we heard from local implementers of OpenStack like Fujitsu, Yahoo! Japan, and - from just across the water - SK Telecom and Huawei.
In keynotes, case studies, and deep-dive technical sessions, attendees learned what worked, debated where to go next, and considered the project's complicated relationship to containers, software-defined networks, the giants of the public cloud, and more.
The hordes gathered in Las Vegas this week, for Amazon's latest re:Invent show. Over 18,000 individuals queued to get into sessions, jostled to reach the Oreo Cookie Popcorn (yes, really), and dodged casino-goers to hear from AWS, its partners and its customers. Las Vegas may figure nowhere on my list of favourite places, but the programme of Analyst sessions AWS laid on for earlier in the week definitely justified this trip.
The headline items (the Internet of Things, Business Intelligence, and a Snowball chucked straight at the 'hell' that is the enterprise data centre (think about it)) are much-discussed, but in many ways the more interesting stuff was AWS' continued - quiet, methodical, inexorable - improvement of its current offerings. One by one, enterprise 'reasons' to avoid AWS or its public cloud competitors are being systematically demolished.