In a sector renowned for impenetrable and usually unpronounceable acronyms, plain language terms are always to be welcomed. So 'Cloud Computing' was initially a welcome term, a metaphor for a new vision of joined up computing over the Internet.
Potentially global, cloud computing would at least be above the level of the individual organisation. It came at a time when it was already commonplace for a wide range of applications and services to be delivered from 'the cloud', from Facebook and more recently Twitter as social networking phenomena to Salesforce.com and latterly Google Apps as valuable business tools.
Terms and definitions
But when the terms evolve to 'public cloud' and 'private cloud' as they have done, it really does suggest that we need a better name. In any event, 'term' is a misnomer when it comes to cloud computing because there is little by way of matching definitions or specifications. Salesforce.com is a particularly good example. It has been highly successful to the point of phenomenon, the first business to even begin to match the popularity of social networking services on the Web. But when it began in 1999, it was the first true Web-delivered, software-as-a-service offering and a cloud was just something on the natural horizon.
By and large as far as its loyal users are concerned it is still just that, a valuable software as a service (SaaS) tool, and they will long since have stopped wondering or caring much how or where the computing is done. The advent of Force.com and AppExchange, enabling third party developers to devise and sell complementary applications, most definitely represents the transition to what can be universally accepted as cloud computing in the form of 'platform-as-service' (PaaS).
Naturally enough, those who are in the cloud are inclined to look down from on high at more earth-bound software providers. "At the same time as trying to climb on the bandwagon, the incumbent software industry is trying to hold back the tide," in the opinion of Dr Steve Garnett, EMEA senior vice president of Salesforce.com. "Traditional computing sells you the boxes and the software, making everything work, especially together, is complex so compared to SaaS they are selling you complexity in order to have you buy support and consultancy. On the Web, there are no versions or upgrades and everyone who buys a service has the same functionality and power. Security is just not an issue, because it is looked after globally by world class systems and expertise. Put it this way, Salesforce.com has persuaded the likes of Merrill Lynch and Deutsche Bank that our security is at least a match for and probably better than their own."
Dr Garnett's clear view is that it is the compelling economic arguments that will accelerate the drive to adoption of SaaS, platform-as-a-service and cloud computing as the architecture on which they run. "The technology is proven and improving constantly, for all users, and even large and conservative organisations have accepted the concepts of multi-tenancy in, for example, sharing data centres. That security concern is still there and will quite rightly always be a serious concern. But then financial services organisations have already developed reliable systems for governance, Chinese walls separation of information, regulatory compliance and so on, all to the best standards of their jurisdictions and sector. So the actual computing architecture has become of minimal relevance to those key security requirements."
While Salesforce.com and its Force.com ecosystem have stayed firmly with the business sector, Google is probably the most fervent cloud exponent and evangelist. "As far as we are concerned, there is nothing ‘coming' about cloud computing. It is here today, we have over 1.75 million users and customers, registered domains, and something of the order of 3,000 more a day signing in on for the business and services offered by Google.com and through GoogleApps." Robert Whiteside, Google head of enterprise for these islands and Benelux, is keen to point out that the specific architecture is just not important compared to the availability, flexibility and performance of what we currently categorise as cloud computing.
"I think a good example is Google Moderator, excellent for managing any forum such as a group discussion or a large online event with subject and agenda setting and participant voting. This is immediate democracy, used very recently for one of President Obama's town hall meetings and online audience. There were over three million online voters for some issues and who know how much higher the next one might go. Moderator could handle the scale and spontaneity of that precisely because it is cloud-based."
Whiteside points to the inter-operability of Google and Salesforce.com as an example of the beauty of cloud computing: "We have different targets, overlapping markets and users and the applications and the clients can work together ad hoc because we are in an environment of open standards and data. Individuals and businesses are not tied to long agreements-a year's subscription is the maximum-and essentially you pay for what you need as you use it. Where stored data is involved it can all be safely and securely moved to the next service provider with not much more than a few clicks."
He echoes Dr Garnett's economic argument, pointing out with some relish the typical reduction in annual costs per user from over €200 to about €40 using a full set of business desktop applications from Google as opposed to the established server and PC based vendors. "Whatever the organisation or typical user needs, that is the order of saving. Trinity College, for example, has just moved its 20,000 student body on to Google Apps." In essence, he suggests, there are no arguments against cloud computing, only specific remaining concerns about specific services.
But Google and Salesforce.com are, of course, already in the remote computing and SaaS and PaaS business. They have grown their data centre infrastructure to do what they do and whether it is called cloud computing or XaaS or remote managed services is simply immaterial. "Cloud computing is still a fuzzy concept with too many definitions, all of which are very broad," observes Paul Daugherty, Accenture's chief technology architect. "I'm inclined to paraphrase the science fiction writer who said 'the future is here today, it's just very unevenly distributed.' There are areas in which very large and smart organisations are already getting serious value from what we can readily call cloud applications. There is for example a large high tech company that deployed an enterprise CRM solution in just two weeks from decision based on a tailored Salesforce.com implementation."
"In Accenture right now we are involved in well over 100 projects for potential global deployments," Daugherty says. "Typical enterprise functions would include CRM, HR, supply chain and other. But then there are areas and types of application that are more deeply embedded in organisations, like ERP systems or stock trading, that require very high level of performance, availability and often security. Cloud computing may offer potential solutions to match those requirements but the guaranteed SLAs and many aspects of the technical architecture are not there yet."
On the other hand, he believes that a hybrid model that integrates cloud and legacy on-premise systems is already here and will continue to develop and work, whether any sizeable number of large organisations will ever migrate to all-cloud computing. "There is a lot of early stage activity in developing hybrid and private cloud solutions but relatively very few organisations looking beyond that. Cloud computing is not yet mature enough for mission-critical applications and it follows that the commercial enterprise grade SLAs are just not there." On the other hand, Daugherty is of the 'Think big, start small' school of thought: "Like other IT step changes in the past, just because there seems to be an inordinate amount of hype does not mean cloud computing will blow away. In fact I would suggest that for any organisation it would actually be dangerous to ignore it completely."
Dell is another major IT company with no doubts about the future of cloud computing and the specific Dell Cloud Solutions (DCS) division to meet what it describes as 'hyperscale data centre' needs. Right there at the engineering coal face, Dell sees itself as provisioning the cloud for cloud service vendors with demands for efficiency, low costs and rapid scaling. It also, as Paul Kenny of Dell global consulting services points out, a major cloud service provider itself with its ProManage and ProSupport offering for 24x7 remote IT management. "We look after everything from software licence management to server uptime to distributed device management to laptop data encryption for literally thousands of users. As a cloud-based managed IT service it is itself an example of the flexibility and scalability of cloud solutions at economically attractive price points."
Dell regards cloud as the provisioning of services according to a unified common model, decoupled from the customer's own IT hardware stack. "But in fact there are lots of interpretations and in many respects the varying technical definitions are not really all that significant. In each case the potential customer just asks if the service will do what is needed, reliably and at the right cost level." Kenny goes on to suggest, like others, that the true barriers to universal acceptance are in areas such as regulation and governance as well as legal jurisdictions for data keeping. "But there is a lot of development work going in solving such issues, guaranteeing the integrity of data and applications. The concept of private clouds is of course one strategy to ensure integrity but we already have to manage issues where private and public clouds meet."
Pól MacAonghusa, CTO of the IBM emerging business centre, takes a clear 'no hype, no mystery' view of cloud computing: "It is a very efficient way to deliver very scalable IT services. In practice, there are different levels of both technical and business complexity to manage in doing that. Consumer services are already well established, essentially cheap and cheerful and the users have minimal interest in how it all gets done. Often services start with beta versions at best and mature rapidly through adoption and popularity."
The challenges to business are in the serious applications and suites of business processes like SAP or Oracle Financials or the myriad vertical market applications. "I suspect we are now in the equivalent of mobile computing in the 80s, with a clunky suitcase sized computer, tied to one vendor or service provider and with no roaming," says MacAonghusa. "Now look at what we can do today with pocket devices and 3G phones, computing 24x7, anytime, anywhere. Users happily and quickly roam between carriers and service providers and believe it is their absolute right to be facilitated in doing so. Take the analogy up to 21st century business level and we can expect, for example, a complete workload to migrate from Cloud A to Cloud B on demand. But that is and will be technically feasible long before general corporate acceptance of that level of cloud services or trust in the appropriate service and quality levels."
But MacAonghusa echoes the others in believing that all of that mostly understandable corporate caution is in practical terms beside the point because the development of cloud, SaaS, PaaS, and infrastructure on demand is just proceeding anyway. "Why invest in resources for peak or occasional demand when you could rent an extra bit of infrastructure on a commodity, pay-as-you-use basis?" So for software development capability or power for compute-intensive projects or tasks, he believes, cloud computing can readily deliver on-demand resources with commodity pricing. That is particularly attractive when the requirement is for some IT function or task that is not too heavily involved with internal business processes.
Microsoft was one of the earliest cloud computing advocates, according to Declan Faller, infrastructure products manager, who says it began work on the concept in 2003 when it and other IT majors began talking about 'dynamic infrastructure'. "Microsoft and of others put the building blocks in place towards that goal of a dynamic ICT infrastructure that would be self-managed and self-healing and scale up or down rapidly. The business objective was to remove the burden of investing in and managing infrastructure away from the user to the provider." Many in the industry were sceptical in those early days, he says, because virtualisation was in its early days and it is key to a great deal of any cloud computing solution.
"We reckon cloud is about 80% there in terms of the technology, with a lot of tough stuff still to come. But then the original goal was seen as a 10-year project." The differentiation that has arisen between public and private cloud is valuable, he believes, because it enables cautious organisations to invest in elements of cloud computing that are here today and practical, blended with what they already have and trust. The two key Microsoft product kits are Dynamic DataCenter to build a private cloud and the soon to be publicly released Microsoft Azure, effectively a cloud services platform or operating system on which service providers can build their hosting and management. "From our point of view, the market key is that these are built so that developers and users of SQL, SharePoint, .NET and other Microsoft technology will find there is no great leap required because cloud will just be an incremental step using similar skills. There will be Azure modules to enable developers to extend current capabilities into the cloud and onto the Web."
Faller is happy to concede that the future of cloud computing and dynamic infrastructure is actually far from clear: "Where it goes is still anyone's guess, really. The realisation of how it could benefit organisations will take time, apart from the fairly obvious pay-as-you-go advantages. That in itself is going to drive adoption of occasional services, for example when for some reason transactions begin to scale exponentially for some reason-not very uncommon in the consumer Web world. A need for high performance for compute-intensive tasks for finite periods and occasional use is a real CAPEX challenge to many organisations. Cloud computing could be the answer in areas such as business intelligence within the enterprise or university and scientific research projects."
It is appropriate to give the last word to VMware in a cloud computing forum because there is no question that the key technology enabler of cloud is virtualisation, in data storage, operating systems and applications. "Virtualised data centres already have what can only be termed private clouds," says Chris Hammans, VMware regional director for Ireland and the UK. "They have consolidated their server estates, moved on to automatic provisioning and are gaining the benefits in CAPEX and OPEX. As for client organisations, they are already recognising that cloud is a logical extension of today's hosting of infrastructure resources off-premise in third party data centres."
He sees the building of general corporate trust in the cloud as crucial, and as much of a priority as the further development of the technology. A key area is traditional outsourcing, well established and generally trusted. "But there is huge complexity in moving to outsourcing. Cloud architecture offers the true ability to be scale flexibly or to move from service provider to provider with no re-writes of applications-in fact no changes-and no inhibiting lengthy contracts. The clients remain in complete control."
He instances the ability under vSphere, as a cloud OS, to move applications off-premise and into the cloud but seamlessly preserving all of the security and business policies of the enterprise absolutely intact. "That alone takes care of one of the major concerns of I would guess every organisation looking at the possible benefits of cloud. If your security and data integrity are guaranteed, together with auditable regulatory compliance, then a major obstacle to running applications off-premise is removed. The same applies to constraints such as, for example, the legal prohibition against holding data outside of certain jurisdictions."
Perhaps it should be said: all of this suggests that one of the major obstacles to the progress of cloud computing at this stage may well be the term itself. It would certainly not be the first time an IT term was offputting to the non-technical market. 'Remote managed services', becoming over time just 'managed services'? Suggestions welcome.