Debunking the Mist

Eduserv’s annual symposium cleared up ‘cloud’ misconceptions and floated a new way of thinking about computing in higher education. Damon Jones reports from the event.

Cloud computing is a frequently bandied, nebulous buzzword – but what does it actually mean – and what will its imminent proliferation entail for HE’s policymakers and IT infrastructure?  

This was the theme of Eduserv’s annual symposium, which took place on 12 May 2011. Gathering at the Royal College of Physicians in London, ten leading researchers in the field assembled to impart their experiences to delegates – and an online audience, for whom the event was streamed live. Drawing from luminaries within the sector and without – including representatives from JISC, Eduserv and the Belfast e-Science Centre – the event exploded some of the murky myths surrounding cloud computing, exploring the rival benefits of public and private cloud provision in HE-specific contexts.

The Morning

With his popular keynote address, ‘Situation Normal’, Simon Wardley of the Leading Edge Forum – assisted by an amusing panoply of ‘lolcat’ graphics – defined the phenomenon, and summarised the social, technological and commercial factors that have led to its emergence.

Puncturing the pompous definition of cloud computing as “an evolving paradigm”, Wardley asserted that cloud services were analogous to an IT ‘national grid’: and merely reflected the migration of key utility services to a new form of delivery. Prompted by attitude changes and well defined operations, amongst other catalysts, less competitive IT functions – such as payroll and HR – are now considered prime areas for cloud servicing, which can provide these operations at reduced cost, leaving businesses free to concentrate on core profit areas. Comparing the cutting edge IT sector with the emergence of the standardised nut and bolt of the industrial revolution, the researcher pointed out that, for the cloud to fully take off, interchangeable systems are a prerequisite.

Interoperability, facilitated by “open source” code – free to all – is a commodity many organisations will be reluctant to relinquish but, according to Wardley, is vital. A refusal to adopt this practice stifles innovation, preventing many potential users from entering the marketplace. As an analogy, the speaker cited the example of the mass media – which no longer has a monopoly on communication, having been surpassed by digital systems which have democratised the sphere. To Wardley, the expense of private clouds is equally prohibitive – and instantly precludes the choices competing public providers may soon offer. Business pressure is such that most HE providers will ultimately feel compelled to keep up with the public cloud trend, and, as he expressed in uniquely feline terms, there’s “no point in turning up to the catfight with a snazzy rifle, if everyone else has got tanks.”

De-misting a significant misconception enveloping the cloud, he asserted that subscription to on-demand systems will not necessarily reduce IT spends. Despite the bountiful resources offered, extra capacity inevitably results in extra usage. The cloud will become increasingly integral, regardless of this false economy – but managing the way it’s used poses formidable challenges. Decision makers in business must now choose between “survival today” – and successful commodification – versus “survival tomorrow” – the innovations that will secure future profits. “That’s the interesting thing about the cloud,” mused Wardley, “when you strip it naked, it’s more about models of management”. Though the technology isn’t new, the way organisations evolve to use it is. Mention of China’s developing ‘cloud computing valley’ - intended to rival its US-based silicone predecessor – and the £763 billion of economic benefits forecast to be at stake between 2010-2015 made apparent the imperative for HE to prudently evolve, and exploit these new strategies to their optimum.

With this dynamic environment analysed from a range of perspectives, delegates began to develop an enhanced understanding of how shared virtualised platforms are beginning to protect and advance services. This was examined both at the level of infrastructure (storage and computing facilities), enterprise applications (CMS and finance), and in established areas such as email and contact management.

The Afternoon

The afternoon session commenced with a set of four “lightning talks” – punchy, ten-minute presentations. Amongst these, Rachel Bruce (JISC) described the breadth of activity under the UMF banner and the operations of its various strands, whilst Kevin Ashley detailed the active management of long-lived digital resources by the Digital Curation Centre. With mention of commercial possibilities – and the BBC iPlayer – buttressed against discussions of purist academic research in the cloud, it became apparent that questions of ethos were equally as relevant as pragmatic concerns to the attendees.

The closing speaker, Armando Fox – well-respected for his pioneering work in the field – discussed these topics. Adjunct Associate Professor of Electrical Engineering and Computer Science at UC Berkeley, Fox suggested the potential merits of the private cloud to HE, and weighed up the competing benefits of the public alternative in his address, ‘Above the Clouds: A View from Academia.’ Citing his own research experience, Fox pointed out that, whilst they may not save cash, investment in private clouds for cutting-edge experiment can generate valuable monies for institutions.

Contrasting with Wardley’s notions of rapacious “organisational warfare”, Fox pointed out that certain experiments could only be conducted on private clouds or ‘clusters’, which investigated areas such as malware and security. Yet these tests also required the participation of public clouds, since the investigations must be validated on a large scale – and critically, by actual users. Access to thousands of computers for these large scale trials would be impossible, however, if it restricted the university’s own resources. “We have done work here that we couldn’t have done, which, right now, is afforded only by public cloud providers,” he testified.
Fox concurred that public cloud services typically seemed cheaper than private equivalents, when intimidating start up costs, potential redundancy, and the “on demand” economy of the privates were calculated. Outage costs, and the disadvantages potentially affecting users crippled by a private ‘single point of dependency’ were also cited as a concern. Another – hitherto unexplored – cost-saving idea Fox raised was that HE players could bond together, comprising a formidable body which could ‘pre-buy’ services, negotiating lesser rates. Referring to one of the major US cloud providers, Amazon.com, he suggested that the provider’s mark-up may be as high as 100 per cent – a staggering total which an academic combine – or greater competition – might challenge.

Concluding his analysis of the public/private cloud debate, Fox argued for a possible synthesis – and for universities to clarify what their specific IT needs are before making critical choices for their infrastructure. “Sustainable and competitive for academia doesn’t need to be necessarily cheaper,” he opined, vouching for the private cloud. “It just needs to be effective in the sense that the product you’re paying for, and the results you expect to get exceed what you could with the public cloud.”