Just when you thought it was safe to go back in the IT waters, yet another beast emerges from nowhere to potentially baffle and confuse — cloud computing. But is cloud computing a shark or a whale: is it a trend that should strike fear into the hearts of senior IT professionals or one that they should embrace?
It doesn’t seem so very long ago that Software as a Service (SaaS) emerged as a concept, promising the delivery of enterprise applications in a hosted fashion over the Internet. Salesforce.com and its ebullient founder and CEO Mark Benioff surely did more than anyone to evangelise the concept, first with hosted CRM applications and today with a platform on which almost any type of application can be found — and all of course delivered as a hosted application that saves users worrying about hardware and patches.
But that wasn’t the first time that applications had been delivered in a hosted fashion: in the nineties the trend was being advocated by Application Service Providers (ASPs), who acted much like salesforce.com but perhaps struck the market a little too early, and lacked the resources and wherewithal to become serious applications players in their own right.
More recently than SaaS though, the industry was deluged with talk of Web 2.0: the second coming of the Internet. A trend that was a little hard to clearly separate from Web 1.0, the Web 2.0 evangelists talked about the importance of end user involvement, rich user interfaces and more sophisticated web-based applications.
Google, Yahoo and even online retailer Amazon led the charge, with more impressive web-based capabilities. They also made web services available so anyone with the right skills could create a ‘mashup’ application that usually remained hosted on Google, Yahoo or Amazon’s web platforms somewhere. At some point — and it seems no one really knows exactly when it was — people started describing this new ecosystem of relatively open, accessible applications and other hosted data sources as ‘cloud computing’.
So what exactly is cloud computing? Wikipedia offers this definition: “Cloud computing means Internet (‘Cloud’) based development and use of computer technology (‘Computing’). It is a style of computing where IT-related capabilities are provided ‘as a service’, allowing users to access technology-enabled services ‘in the cloud’ without knowledge of, expertise with, or control over the technology infrastructure that supports them.”
If you feel that the term doesn’t sound particularly different from some of the trends that preceded it — SaaS, utility computing, grid computing, on-demand computing and so on — you would not be alone. But that is not to say that it should be treated with so much scepticism that it is ignored altogether. As William Fellows, principal analyst at The 451 Group wrote in a recent report on cloud computing, “Cloud is the name for a real trend (utility computing 2.0), as well as an accompanying bandwagon. It’s the ‘baby and bathwater’ principle: We believe that there are good reasons not to dismiss the trend itself, despite the hype.”
The 451 Group defines cloud computing as IT as a service — as opposed to Software as a Service (SaaS): “The cloud is IT, presented as a service to the user, delivered by virtualised resources that are independent of location,” said Fellows.
It’s thought that people started using the term ‘cloud computing’ around early 2007, but the hype of which Fellows talks has really gathered a head of steam this year. It’s being driven mostly by companies like Amazon, Google, salesforce.com and Yahoo, as well as more traditional vendors including Hewlett Packard, IBM, Intel, Sun and Microsoft.
So is it a ‘real’ concept or more of the same old marketing spiel? Annrai O’Toole was CEO of Cape Clear, an enterprise service bus (ESB) company that optimised its integration platform to enable what it called ‘integration on demand’. It was recently acquired by Workday, a business applications on demand vendor, where O’Toole is now VP of integration. What does he make of the term ‘cloud computing’? “At first I was sceptical,” he told CBR when we caught up with him last month. “But then I heard Dana Gardner [president and principal analyst at Interarbor Solutions] give his whole shtick on cloud computing and I bought it. I definitely buy it now.”
According to O’Toole, the concept is becoming more clearly defined as time passes, and the distinction between cloud computing and its precursors like SaaS and on-demand computing is becoming more clear. While SaaS was software as a service, and on-demand computing from the likes of IBM was really processing on-demand, cloud computing is wider-reaching: it is IT as a service. It could include not just applications but processing power, storage, connectivity, collaboration, extensibility, configurability, application development tools and so on.
O’Toole draws another distinction between SaaS and cloud computing: “There is a subtle but important difference for CIOs,” he says. “This software is far more configurable. SaaS was customisable, but there is configurability and extensibility with cloud computing that was simply not possible with earlier generations of SaaS.”
O’Toole says that a fundamental requirement of cloud computing is some sort of brokering in ‘the cloud’. “Take Amazon,” he says. “It opened up its platform so retailers could plug into it. It enabled people to customise the platform, thanks to modular components. Applications or modules can be enhanced, customised, configured — all plugged into the Amazon EC2 or S3 platforms. SaaS is just an application on the end of a URL: you can’t plug anything in to it.”
Not everyone is convinced it is markedly different from its precursors. Butler Group research director Tim Jennings wrote in a recent opinion piece: “I believe that the leap forward to cloud computing is not as large as it may seem at first sight. For large organisations today the data centre is often located at a remote site, accessed over network connections; it is also becoming highly virtualised, breaking down the direct relationship between particular applications, servers, and storage.”
“There is relatively little difference between this and the cloud computing model,” Jennings said, “and I believe we will see a blending of the two models, as organisations both develop their own internal cloud infrastructures, and supplement existing capacity with external cloud services.”
Which brings us to two very important questions: is cloud computing ready for the enterprise, and for that matter, are enterprises ready for cloud computing?
The first question is relatively easy to answer. There are few companies claiming to offer real cloud computing capabilities to the enterprise today, and those that do are usually really doing SaaS or storage or processing on-demand. For sure, the likes of Amazon, Google, AppNexus, Joyent and GoGrid offer varying degrees of cloud computing but these are aimed really at consumers or small to medium enterprises. It’s unlikely larger enterprises would see them as a viable alternative to their in-house IT infrastructure just yet. IBM, Sun, Intel and Microsoft have more enterprise-ready credentials, but even they cannot claim to have reached the point where they can offer true cloud computing.
But they are definitely getting there. In July Hewlett-Packard, Intel, and Yahoo entered into a partnership to create new research and development centres for cloud computing. The companies will set up six test facilities called Cloud Computing Test Beds for cloud computing infrastructure research, with up to 4,000 processor cores each.
In August, IBM announced plans to expand its cloud computing capabilities by investing nearly $400m to set up two data centres in the US and Japan. The centres are part of IBM’s Project Big Green initiative aimed at increasing energy efficiency in the data centre.
3tera, which already offers various on-demand and hosting capabilities, has announced what it calls Cloudware, the kind of brokering platform O’Toole talked about: “Cloudware is an architecture intended to provide an open framework allowing the development of a cloud computing environment that’s rigorous enough to take on any web or enterprise application,” the company says. It says it will be rolling out the new cloud computing capability in the next 12 to 24 months, adding the likes of support for Solaris and Windows to existing support for Linux; choice of multiple data centres worldwide; pre-built MySQL clusters; database replication appliances and NAS integration with third party storage solutions.
In fairness, cloud computing for the enterprise is not with us just yet. So how about the next question — is the enterprise ready for cloud computing? That is a little trickier to answer, and like so many questions in IT, is probably best summed up as, ‘it depends’.
Some enterprises have already seen the benefits of SaaS and perhaps on-demand computing, so for them cloud computing is the next logical step. With its ability to save companies electricity costs (the data centre is effectively outsourced) and reduce IT management, configuration and maintenance headaches, it’s got some compelling advantages. It also means paying for IT resources rateably, instead of making capital investments and then depreciating them.
For others, it will be a leap too far. As Sybase CEO John Chen told us when we met him in London last month, “There’s always a customer preference thing with IT. A lot of people want to buy it [their own IT infrastructure], own it, treat it as a capital expense and depreciate it over time.”
Chen says the company has various cloud computing capabilities to pursue thanks to the likes of its iAnywhere mobile messaging and management platform, analytics, integration and data warehousing. “You still need a database and messaging in the cloud, so this is all goodness to us,” he says.
Of course, as with any new paradigm in IT, there are potential drawbacks to work past. “Service levels, security, and performance are all areas where customers will need reassurance and clear evidence of adequate resilience,” according to Butler Group’s Jennings.
The 451 Group’s Fellows sees even more potential stumbling blocks: “Security, red tape, fate sharing, legacy infrastructure, connectivity, offline access, performance, volume, SLA management, job security and territorialism, to name a few.” Against this backdrop, only ‘bleeding edge’ enterprises are likely to be ready for cloud computing today.
More and more IT vendors are working out what their strategy should be to embrace cloud computing. As for enterprises, many are investigating the concept. As is so often the case, it is unlikely cloud computing will replace traditional IT, just as SaaS has failed to eradicate client-server applications. But as the vendors build out their cloud computing capabilities, enterprises would be shrewd to keep an eye on developments. Bewildering, vacuous or revolutionary, cloud computing today can perhaps best be summed up by this comment from analyst firm Gartner Group: “The very confusion and contradiction that surrounds the term ‘cloud computing’ signifies its potential to change the status quo in the IT market.”