Following CBR’s cover feature on Cloud Computing earlier in the year, Andy Mulholland, global chief technology officer at Capgemini, takes a closer look at the implications of this new paradigm for enterprise IT.
The way we do business is changing. In the 21st Century, organisations no longer operate in a fixed, predictable environment, so neither can IT services expect to be provisioned in the same way.
IT management is expected to deliver the capabilities and support for this continually changing landscape. Every organisation’s business model and IT architecture must be able to cope with rapid change and be able to support flexible relationships with customers, suppliers and partners; the challenge is as much across the firewall as within the firewall; as much as a shared external capability with customers and suppliers as a fixed internal resource.
To meet this challenge, IT directors must take a fresh approach — moving away from the existing centralised, static application architecture towards the flexible, agile approach based around ‘services’. Starting by using Service-Oriented Architecture (SOA) to start to deliver business aligned services internally, and to understand how to use services externally supplied by the cloud computing environment.
Cloud computing is a major new approach to delivering IT infrastructure services that will increasingly become essential to the business, because of its capability to deal with rapid change in external markets.
Cloud computing has become ubiquitous over the past twelve months, peaking with a ten-fold increase in news articles in October; but what is the reality behind the hype? Is cloud computing simply grid computing in new clothing, or does it really herald a ‘new era’ in computing – an era where technology and business services are essentially the same thing?
Analysts Gartner define cloud computing in their annual survey of technology hype as a style of computing where massively scalable IT-related capabilities are provided “as a service” using internet technologies to multiple external customers. But multiple definitions exist and how can a CIO distinguish between hype and real hope?
Large internet and technology companies are pressing forward to deliver information and software over the net. Whilst IBM, HP, and Sun can be recognised as early leaders, competing for the supply of the technology services, recent announcements have brought in players such as Oracle and EMC. Microsoft recently followed the lead of Salesforce.com, by releasing its own platform for building cloud-based applications, called Windows Azure. David Thompson, corporate vice president of Microsoft Online, announced that the company would eventually move all its enterprise software into the ‘cloud’.
Infrastructure cost play
There are three main elements which make up the invisible infrastructure of cloud computing as a platform. Everything – from raw computer power to storage capabilities – sits on the bottom layer. These elements are all tied together and delivered as a single integrated entity under sophisticated policy and event management that embraces everything we know from today’s enterprise IT, together with new services to deal with complex issues such as security and authentication.
On top of this sits the platform through which the technology is delivered to be consumed; or rather it may be a series of platforms from different technology and business services vendors. Platforms existing to provision their specific services to be used by others and that means they need to be ‘open’ terms of published APIs. Some of the best known platforms today offer widely used generic capabilities such as Google Maps, or Facebook communities, and are often referred to as ‘cloud services’.
Google, for example, is supporting the use of its platform with a range of services by an unknown number of people, or services, in an unknown time frame – no conventional IT system could cope with this. The infinite flexibility of the resources that Google can make available from their range of global data centres by pooling capacity under cloud computing is a fundamental part of being able to operate their platform.
The final layer contains the ‘business services’ meaning those services that are created and deployed to be uniquely differentiating to an enterprise and its customers, or suppliers. Something such as a logistics company building services to locate your parcel with particular benefits that will be unique to their company, but it will be built over a map of the neighbourhood coming from the Google platform as a set of services to be consumed, and due to the external, and unpredictable demand, all of this will be supported by cloud computing technology capacity from a further supplier such as Amazon and its Elastic Compute Cloud, EC2.
Observers have been interested to see that Microsoft has now joined in with cloud computing (see CBR’s Online page this month) and it is worth examining their offering as an illustration of how technology vendors see their path into cloud computing. The Windows Azure Services Platform is an internet-scale cloud computing and services platform hosted in Microsoft data centres that provides a range of functionality to build applications that span from consumer web to enterprise scenarios. This covers both developer tools to integrate the Microsoft cloud services and to build unique business services above.
Windows Azure is essentially a bundle of Microsoft services covering everything from the ‘Live Suite’, through to SharePoint, and Dynamics CRM. It also features .Net Services and SQL Services and integration with Visual Studio to make a consistent development environment.
The platform effectively ties Microsoft environments with non-Microsoft environments together and offers standard interfaces to external non Microsoft environments, which leads Microsoft to dub it an ‘open platform’. This is an important point. For a fully fledged cloud approach, the extent to which the platform allows the services to be consumed by layers of business services in a unique combination with other services, defines the extent to which it really is an ‘open’ platform.
Cloud computing business models
Much of what is in use as generic services on the internet and web is in fact already based on using some degree of the cloud computing model, and is linked to other changes in the way that business value is created and paid for. There is much speculation and debate about how long ‘free’ web services which are paid for by advertising, can survive. However, ‘free’ as a business model is perhaps already being changed as the successful players have used ‘free’ to create new business value that can be sold by one of two models known as ‘freemium’ and ‘complementary’.
‘Freemium’ in essence means giving something away for free to create a market and demand, then charging those who are looking for more than the basic service. Adobe is a great example of this. Free downloads of applications such as Flash, or PDF, have created a global market standardised around their technology.
The question is just how many areas this can work in. Some of the best known success stories, such as Flickr and LinkedIn, started as advertising-supported and migrated into ‘freemium’ funding. So the cloud computing model in these cases is already making these services available as cloud services running on a platform, and in some cases with premium charging.
The real money, however, both as an investment to get started and in terms of potential to build an unassailable position, belongs with the ‘complementary’ model, and Google is a perfect example of this. Google is an example of advertising funding but the Google franchise is starting to reap income from other sources.
Google has, and is, using free services around search to build a huge information base, which it can sell not just around advertising, but as a widening range of complimentary services to commercial users. The key question is how to make the various elements of a ‘complementary’ business model accessible to be consumed, by the ‘net economy, and this again means adopting the use of cloud computing services delivered from a web accessible platform.
Suitability, and risk, for the enterprise
We need to think of a new term for the provision of the more complex set of capabilities that extend outside our organisation as well as inside and are the basis for an increasing amount of our personal work, and our enterprises capabilities to do business. These services are more, much more than the old definition of ‘infrastructure’.
We should perhaps think of it as an ‘invisible infostructure’ defined as a common and shared environment which uses cloud computing to combine a wide range of hardware with software and management services – including all facilities to exchange and process (inter) company information – as a commoditised, preferably almost invisible utility. Eventually, even some non differentiating core business services may merge into it, creating a true ‘business infostructure’. An invisible infrastructure already exists in the internet and the web with their ubiquitous standardised capabilities shared by all; so the extension of this in cloud computing is a natural evolutionary step.
Many enterprises have learnt how to use the internet and the web to create a wide range of business valuable capabilities. Learning how to use cloud computing as a model to still further develop this through partnering is their next step. Those who fail to learn how to do so will find themselves the last choice in any new business relationship as the prize will go to those who can use the technology partner easily and well.
There is a long way to go before the current generation of enterprise computers and applications will embrace cloud computing. It is more likely that cloud computing will develop most rapidly to supporting web based systems, with the conventional data centre continuing to support our current generation of enterprise applications. However that is not to recognise the wide range of tasks that exist and demand unpredictable support from time to time such as data cleansing, for these using cloud computing technology services will be an obvious effective move.
The very real risk for today’s CIO is that business and technical drivers will increase the need to change the approach to delivering projects, but that their organisation will fail to recognise the need to act above the level of individual projects and put in place a systematic approach to switching to SOA, and a roadmap to cloud computing.
Capgemini strongly recommends that CIOs begin considering how they will operate in a SOA environment. This doesn’t mean ripping out your architecture and starting again. It means building a roadmap of how you will make the shift to SOA and how much of that will be built internally, and how much would be better sourced from third parties. Luckily, moving to SOA as your enterprise application architecture is not a side-road – effectively utilising SOA is a requirement for effective use of cloud computing.
Read our recent CBR cover feature, Is The Enterprise Ready for Cloud Computing, at tinyurl.com/57cp39.