Cloud Zone is brought to you in partnership with:

Cloud Computing/ Big Data/ Open Data/ Linked Data Consultant & Analyst. Podcasting with tech execs. Analyst for GigaOM Pro. Based in the UK, working world-wide. Paul is a DZone MVB and is not an employee of DZone and has posted 34 posts at DZone. You can read more from them at their website. View Full User Profile

Big Cloud, Little Cloud, Local Cloud, Global Cloud

07.11.2012
| 2913 views |
  • submit to reddit

Amazon’s globe-encircling cloud infrastructure is compelling to many. From Virginia to California, from Ireland to Singapore, and from Japan to Brazil; wherever you find yourself there’s a local instance of the same familiar set of services. And, in all likelihood, Australia will soon be added to the list. For those primarily interested in just serving both Europe and the US, the list of options grows to include Rackspace, GoGrid, CloudSigma and a few others. And yet, despite the buying power and increasing ubiquity of these larger players, there seems to be plenty of space left for smaller entrants. For prospective customers only concerned with a single country or region, for example, the choices are almost too many to count, and choosing between them becomes a complex and multi-faceted affair.

Take Brightbox, for example. As my family know all too well, those pesky timezones mean that many of my evenings are punctuated with calls to or from the States, where so much of the innovation in this sector continues to take root and grow. Either that, or I’m creeping out of a sleeping house to catch early trains for the 150 mile journey south to London. It was therefore refreshing to talk to someone in this industry whose offices are only 50 miles away in the UK city of Leeds.

Established back in 2005 as a Ruby shop capable of hosting apps on dedicated hardware, Brightbox has evolved to place increasing emphasis upon the provision of infrastructure. In 2010, the company began to seriously explore the possibility of offering a generic cloud infrastructure environment. This was in the days before OpenStack, but Eucalyptus existed and was attracting interest. But according to Brightbox co-founder Jeremy Jarvis, 2010′s Eucalyptus lacked some key resilience features (load balancing, multi-data centre capabilities, etc) that the team believed were critical… so they built their own system from the ground up. And, at the end of September last year, Brightbox Cloud entered general availability.

The Brightbox cloud operates out of two UK data centres, with planning underway for a third. Both data centres are now owned and operated by Telecity, which acquired the two independent data centre providers with whom Brightbox had launched. Brightbox owns the racks and (Dell) hardware, and also ensures provision of redundant network access into the data centres. Customers are predominantly drawn from across Europe, but Jarvis says he’s seeing some customers coming from as far away as Australia, New Zealand, and Brazil. The company sees its potential growth through eventual expansion to data centres on the European mainland, but Jarvis says he’s “much less interested in setting up yet another” North American operation. The company is profitable, employs ten staff, and is seeing steady growth in usage.

Brightbox does not (yet) offer a web management console, but Jarvis describes this as a conscious decision and also something of an asset. The company has instead focused their attention upon crafting a rich, capable and intuitive API (and associated command-line interface). According to Jarvis, the developers that the company tends to target have responded favourably to the API, describing it as “nice” and “more consistent” than the various APIs offered by Amazon’s growing suite of services.

A focus upon developers (and the growing Dev/Ops movement) was also a strategic decision, and Jarvis cites examples in which ‘mere developers’ have proved instrumental in securing significant contracts with Brightbox from their employers. Corporate purchasing processes may, finally, be evolving. Despite the success of Platform as a Service (PaaS) offerings such as Heroku and EngineYard, Jarvis also suggests that Brightbox is seeing growing evidence that developers are seeking more fine-grained control over infrastructure than PaaS typically offers. A platform abstracts the underlying complexity of infrastructure, making it easier for application builders to focus upon creating the specific services they wish to provide. But abstractions typically require compromises, with configuration decisions being made for everyone on the platform on the basis of ‘normal’ requirements. For developers with non-normal requirements (and they may actually be the majority of users), IaaS is more likely to offer the fine-grained control that they need.

But the cloud infrastructure world has come a long way since Brightbox began planning their product two years ago. OpenStack has arrived, and (despite a growing body of nay-sayers) is credible. Rackspace, HP, and others are on the cusp of delivering real clouds to real customers on the back of its codeline. Eucalyptus appears to have turned a corner, and pulled back from a brink that I (and others) saw them teetering on the edge of. Canonical’s marriage of Ubuntu to OpenStack is now just one of several ways to get the same cloud code, capabilities and apis onto machines running inside your own data centre. Amazon just keeps on doing what Amazon does, incrementally adding features, cutting prices, and becoming ever-harder to not choose.

In that world, surely a little cloud provider operating their own bespoke solution from the wrong end of the Leeds-London railway line, in a country on the wrong side of both the English Channel and the Atlantic Ocean has no hope? Surely they should just pack up shop, and either adopt OpenStack/Eucalyptus fast… or find a new line of work?

Jeremy Jarvis disagrees, vehemently. And he’s not alone. Look at Edinburgh’s Flexiant, Glasgow’s SymetriQ, and a whole host of other companies that have built their own solutions from nothing. Others, of course, have recognised the value in taking OpenStack, Eucalyptus, Microsoft’s Windows Azure, and similarly established technologies, and making them their own. Look at Eduserv’s Swindon data centre, or the hosted desktops from East Yorkshire’s own GoCloud.

So how can Brightbox (or Flexiant, or SymetriQ, or any of the other non-conformers) compete? How, indeed, can they survive? Jarvis suggests that “Buy British” continues to carry weight here. Even without raising the scary (and often, actually, wholly irrelevant) spectre of PATRIOT Act-powered snooping, a significant proportion of UK (or European) companies like the idea of buying services from UK (or European) suppliers. They like that the documentation is spelled correctly. They like that telephone support is (more or less) in their timezone. They like that the development team shows up at local events, and that it’s a person buying the drinks in the bar, rather than the disembodied marketing budget of some far-off corporation.

Purchasing decisions for something like cloud infrastructure are complicated. Often, they’re probably quite illogical. Price isn’t always the deciding factor (and even if it were, smaller providers like Brightbox aren’t ridiculously expensive in comparison to their larger competitors). Having been the friendly face at a local developer event might swing that big contract. Or having a “nice” API. Or implementing niche features in firewalls, or networking, or port forwarding might each grab the attention — and loyalty — of specific sectors of the long tail. Tesco might sell ‘everything’ to ‘everyone,’ but we still have room in our lives for the SPAR corner shop, and for the upscale deli with the nice cakes.

The same’s true in the cloud, although I can’t help feeling that we’re going to see quite a rapid decline in entirely new cloud infrastructures as the next generation of niche cloud boutiques take OpenStack or Eucalyptus and mould them to their requirements. They probably won’t be making those decisions because (like we Clouderati) they agonise endlessly about interoperability or portability or the de facto standard of the Amazon Web Services stack. For most of their SME customers, those things simply don’t matter. Instead, they’ll be adopting OpenStack or Eucalyptus because the grunt work (and the marketing) has been done. It simply costs less to take something off the shelf than to develop it yourself from scratch. But for companies like Brightbox, where that investment has already been made? Well, for them there may still be plenty of prospective customers out there.

Published at DZone with permission of Paul Miller, author and DZone MVB. (source)

(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)