What is cloud hosting?

in the context of service hosting is one step up from 'standard' shared hosting. Rather than have services based on a single physical machine (with constraints on RAM, processing power, disc size and so on), services run on several machines. These machines don't even have to be in the same location. As long as the interconnections between disparate sites are fast enough (gigabit ethernet is usually considered to be the minimum speed) clouds can be spread over any area. As computing power can be harnessed from several sources, there is a great deal of flexibility available both to the service provider and the end user.

The distribution of computing tasks is managed, invisibly to the end-user, by software called load balancing software. There are many vendors of this type of software, with sources ranging from free open-source solutions to proprietary, more expensive versions. It is this software that allocates processing tasks to a number of resources. For the data centre, load-balancing has a number of advantages:

• single or multiple physical systems removed, added, upgraded or taken on- and off-line with no disruption to the customer. As and when systems are removed or added, the load balancing software can allocate replacements for missing systems, or incorporate new additions into the whole.
• there is no need to have any system not utilised. In traditional computing environments, it was possible to have systems left under used, or not used at all. In a cloud environment, every system can be available for load balancing software to utilise.
• server hosting providers can quickly allocate resources to customers according to their requirements. Indeed resources can be granted simply by easy configuration of load balancing software, so the provider can offer very short-term changes in a service's power allocation, according to demand - changes as short as a few hours.

For the end user, cloud computing has several advantages

• sureness of uptime. Because multiple systems are running a service, any system failure will not have catastrophic impact on the service's provision. In fact, there may not even be some very short-term drop as the provider's load balancing software should be configured to automatically replace failing systems on the fly.
• scalability. Systems can be scaled up or down in terms of their power very quickly and easily. This allows system changes to respond to, for instance, seasonal change, or allow swift response to market pressures. As mentioned above, changed in service power can be instigated very easily, so changes in required power that are only a few hours long are a real viability.
• multiple platform technologies. If a service requires previously incompatible technologies such as PHP and ASP.net, these resources can be provided from different parts of the cloud but served in a unified offering.