Image: NASA

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.

Related articles

• Eucalyptus grabs $30M from IVP to push the open-source private cloud forward (venturebeat.com)
• Announcing Brightbox Cloud – the UK’s first true IaaS platform! (brightbox.co.uk)
• Platform as a service ready to rev up, says Gartner (zdnet.com)
• Hosting providers aim to challenge cloud leaders (zdnet.com)
• Joyent Brings Its Public Cloud To Europe (techweekeurope.co.uk)
• Rackspace launches OpenStack-powered next-gen public cloud (venturebeat.com)
• Ubuntu Cloud Portal: Brightbox 12.04 daily images now available, discounts for testers and Ubuntu members (cloud.ubuntu.com)
• Ubuntu Linux heads to the clouds (zdnet.com)

The company was founded in 2003, and is currently profitable with some thirty employees, no debt and no external financing.

According to Malviya, the team set out to create a multi-tenant Platform optimised for the easy creation of web-based applications, but their first public offering was a SaaS CRM application devoted to the media industry. That application is still running today, and is used by more than 150 enterprise customers.

The Platform followed in September 2007, exclusively as a hosted offering from LongJump’s partner data centres, and Malviya perceives it as in direct competition to Salesforce.com’s Force.com. With extensive support for common protocols and communications specifications such as REST, SOAP, Java and XML, and a common security framework throughout, Malviya argues that an application Platform such as the one his company offers is important in overcoming what he describes as the

“Enterprise tendency to develop one-off apps… and then expect IT to support them.”

He argues that business users are able to take a degree of control over their own application needs, whilst reassuring the CIO and IT Team that everything is running on top of a single set of secure infrastructure over which they have control.

Today, that same Platform is being made available on an annual subscription basis for local installation behind the firewall, as well as being offered to a range of third party hosting companies whom Malviya hopes will brand and resell the software to their own customers. According to Malviya, two (unnamed) service providers have already agreed to resell the Platform, and details will no doubt be forthcoming as they get up to speed.

As the LongJump corporate site notes,

“LongJump BAP for Enterprises addresses a major gap in next generation application platforms designed solely for the public web. Information that is sensitive or subject to regulatory issues is relegated to sitting on a custom application built internally, adding to a level of complexity and management that neither scales nor adapts well with changing requirements. Because LongJump BAP for Enterprises sits inside your own datacenter, its applications and platform are ideal for organizations tasked with managing [sensitive] information”

And, crucially,

“In addition, applications can be packaged for migration from one LongJump deployed instance to another. For example, development can take place in a cloud deployment and be deployed in production in a corporation’s data center.”

This will, of course, be crucial in allowing flexible movement between different environments as requirements, capacity and resources change.

Related articles by Zemanta

• Marketers Beware: Even Dilbert Says Everything is a Platform (marklogic.blogspot.com)
• Mashups Quickly Emerging through PaaS (programmableweb.com)
• Developers are bullish on PaaS (infoworld.com)
• PaaS risks (accmanpro.com)
• PaaS, Trusting Beyond Its Initial Hype (cloudave.com)
• Salesforce.com: Pondering the next 10 years (news.cnet.com)
• Understanding SaaS business models in conversation with Adam Gross of Salesforce.com (cloudofdata.com)

Those podcasts have tended to focus specifically on the Semantic Web, but we’re keen to broaden that scope a bit to tackle a wider set of the issues facing those seeking to run data-powered businesses out on the Web. SaaS, PaaS, Cloud Computing and the rest are clearly vital pieces in that particular jigsaw.

So watch out for a broader set of topics in those podcasts, and thanks to Duane for his time in getting things kicked off yesterday. Do have a listen, and let us know what you think…

Related articles by Zemanta

• As a Service: The many faces of the cloud
• Any aaSes in your organisation?

Have a read, and see what you think.

I spoke to Sinclair recently, ahead of today’s announcement of their SaaSGrid offering, and there’s plenty more to share from that conversation when I get to it in my task list!

Related articles by Zemanta

• Visual Map of the Cloud Computing/SaaS/PaaS Markets: September 2008 Update
• *aaS is getting out of hand
• Understanding the Cloud Computing/SaaS/PaaS markets: a Map of the Players in the Industry
• Silliness-as-a-Service
• As a Service: The many faces of the cloud
• SaaS Soup: Navigating the “as a Service” Acronyms: CaaS, DaaS, DBaaS, PaaS, SaaS, XaaS