The latest fashionable idea, ‘cloud computing’, reflects a reality: the Internet has come of age. But this will only represent a real opportunity for business if what is on offer takes full account of the very significant requirements of business computing.
IT has always advanced along with the flow of fashions and ideas that claim to be ‘revolutionary’. The latest of these phenomena, ‘cloud computing’, builds on fundamental trends that have undeniably been establishing themselves for several years now: the growth in outsourcing of IT tools and processes, the development of applications accessible via the Internet (ASP, Web Services, SaaS…), the standardization and virtualization of underlying infrastructures, and the implementation of pay-per-use principles.
It is the combination of all these factors that has now led Amazon to rent out processing capacity (with its EC2 ‘Elastic Compute Cloud’ service) and data storage (S3). It has led Google to develop a Web browser (Chrome) designed as a portal for all its applications (Google Apps, Gears...) and to offer a mobile phone platform (Android) that will compete directly with Apple’s iPhone. And it has led SalesForce.com to extend its on-demand CRM (customer relationship management) software to a myriad of partners. ‘Cloud computing’ is all about being able to draw on software and hardware resources available via the Web, depending entirely on need. Or, to repeat the analogy made by Nick Carr in his book The Big Switch, to use the network to ‘supply’ our processing requirements from massive central ‘power plants’ that deliver ‘computing power’ on tap, in the same way that power stations produce electricity.
In many ways this is a very exciting and appropriate vision. Most importantly, it recognizes that the Internet has become a real application platform offering an abundant range of tools. It also highlights the enormous potential offered by massive resources that are standard, flexible and homogeneous. On the other hand, it is a vision which seems to reduce computing to the level of individual usage, or at most, to the smallest of structures. For the time being, it fails to recognize the significant constraints that weigh down on enterprise computing: security, availability, performance, heterogeneous systems integration, robustness of processes... Even a cursory examination of the challenges that IT Departments face every day is enough to confirm that not everything is ‘cloud-computable’, despite what the most fervent supporters of the idea would have us believe.
‘Cloud computing’ really does offer some very interesting prospects for businesses provided, however, that it seeks to properly adapt itself to organizations’ needs and the way they want to use it. It is very likely that the first of these adaptations will take the form of a segmentation of the available offerings. The pleasant but idealistic concept of a gigantic, universal computing resource will give way to more specialized computing plants, designed by experts for specific markets and targeted needs.
Today, it is possible to access office automation, messaging, contact management and CRM applications via ‘cloud computing’. In the near future, we can envisage that these specialized computing plants will cover many other kinds of applications and tools (Business Intelligence, Computer-Aided Design...) or even offer resources that combine hardware and software infrastructures such as, for example, storage subsystems linked to archiving solutions, or High-Performance Computing units supporting computer modeling solutions. Above all, this will require a high level of technological expertise, because it will involve having the wherewithal to establish large-scale specialized computing plants. These will need to be capable of automating the whole range of operational processes involved and guaranteeing the levels of service that users expect; designed to effectively manage heterogeneous systems and adapt to the wide diversity of customer demands, and finally, set up to provide the optimum, scalable topology to deliver flexibility, effectively allocate the available power, and minimize carbon footprints. All areas that Bull has concentrated on in the development of its Bio Data Centers™.
But in order to serve businesses effectively, however powerful they may be, the computing plants of the future cannot afford to be just centers specializing in data processing or data management. These specialist agglomerations of technological resources will also have to offer an extremely comprehensive range of services. In the first instance, these services will need to include the capacity to develop and/or integrate a huge library of applications, as well as all the techniques and methodological and legal tools that will allow them to be used under business-critical conditions (not just access portals, task planning, security policies, tariffs and billing conditions, but also disaster recovery procedures, legal archiving, service level guarantees, audit trails…). These services will also need to take the form of a genuine business support resource, if they are to be capable of translating each new requirement into an operational solution, and to facilitate the transition to new ways of using computing resources that may involve turning some conventions on their head!
The standardization of computing resources, more widespread access to high bandwidth networks and the development of Service-Oriented Architectures (SOAs) are the bedrock of ‘cloud computing’. You can bet that in the near future this term itself will disappear, with the advent of specialized computing plants based around an automated approach to business functions. As ‘Architect of an Open World’™, Bull is actively contributing to this evolution process. And as a long-standing partner of companies and public sector bodies alike, Bull understands how to translate diversity into richness, and knows that even though you may be able to standardize technology platforms, you will never be able to standardize the business needs of every individual organization.