Cloud computing is a major cause of confusion amongst businesses at the moment. Though it's still not clearly defined, many are calling cloud the "next big thing" and the technology has been the subject of much media and vendor attention. It's no wonder many IT managers find it difficult to sort the hype from the reality.
What is Infrastructure as a Service (IaaS)
For the purposes of this article, I will focus on the infrastructure services side of cloud computing, commonly referred to as Infrastructure as a Service (IasS). What is Infrastructure as a Service? IasS is the rental of shared and virtualised computing processing, memory, and data storage for the purpose of running an operating system such as Microsoft Windows Server or Linux. This is in contrast to the hosted software application side of cloud computing, generally known as Software as a Service, which is not relevant to the topic of this discussion.
Cloud computing is suitable for homogenous applications that do not require much customisation and are not subject to strict compliance regulations. These applications are less mission-critical, so businesses can accept a lower level of service and cope with a less available application. Email and storage of archived data or an application testing platform are prime examples, as most organisations could survive without them for an hour or two. In exchange for less reliability and availability, the service is naturally cheaper.
Cloud computing has huge potential, but the need to demonstrate value against business objectives should always be key to its long term adoption.
Michel Robert, Managing Director, of Claranet UK,
Applications unsuitable for the cloud
Applications that are not suitable for provision via cloud computing tend to be those that are absolutely mission-critical, heavily regulated, and require extensive customisation and management. For example, an e-commerce engine for an online retailer must be audited to comply with PCI-DSS standards, which is practically impossible when the server infrastructure is shared with many other companies. Other examples include financial services applications that carry customer data and development platforms used by software houses. It's important to note, however, that many of these applications may be suitable for provision via an outsourcing arrangement with a managed services provider that uses dedicated infrastructure.
What I believe we'll see develop in the near future is a clear distinction between those who already own the assets to provide large-scale cloud computing and those who don't. For example, we at Claranet have always strived to remain asset-light to retain agility and take advantage of what the cloud can do to help customers. So we're looking at the offerings of big players such as Amazon, Google and Microsoft as they evolve, and are talking to them about using or reselling their services as a component of a broader managed service.
SMEs benefit from managed service providers
From a purchaser's perspective, most small and medium-sized enterprises (SMEs) benefit from cloud computing already if they buy online services, but they wouldn't necessarily know it. Some SMEs may also have very specific infrastructure requirements that could be enabled directly in the cloud, but I believe these organisations are in a very small minority. So for infrastructure as a service sold directly to SMEs, the market won't be large. However, the market for managed service providers selling infrastructure services to SMEs and using the cloud for some parts of the infrastructure, the opportunities are huge.
Practical advice for IT managers in SMEs
• First, I cannot stress enough the importance of clarity around the service-level agreement (SLA) and, in particular, how important it is to have an SLA on all IT services, from office to cloud hosting. This ensures that when something goes wrong, no one gets caught finger pointing.
• Many IT managers are attracted by the low price of storage services like Amazon's S3 offering. It certainly does offer good value for the money, but it is purely for storage, not applications, and by looking around you may find better value through bundled services.
• It seems that there is an erroneous view among IT managers that cloud hosting is just a pound-per-hour model. However, don't forget that once you go above one physical machine, it's far more cost effective to have a hybrid cloud or dedicated hosting model where predictable usage is served by the dedicated infrastructure and the occasional burst is managed in the cloud.
• It's important to pick an IT vendor who isn't wedded to one particular technology or cloud infrastructure so that you have more choice. Cloud is still developing and you don't want to be stuck with the wrong technology.
• Lastly, finding a clear, workable definition of cloud computing is important to understanding it and its potential. However, what is most important is how businesses start using cloud in their IT strategies. Ultimately it's a question of what businesses need from their IT function and, in the case of outsourced IT, the level of service required for them to operate and grow. This applies whether it is archived storage in the cloud with no availability guarantee or dedicated hosting and application management guaranteed by one end-to-end SLA.
Cloud computing has huge potential, but the need to demonstrate value against business objectives should always be key to its long-term adoption.
Michel Robert is the managing director of vendor-independent managed service provider
This was first published in December 2009