An Introduction to Cloud Servers and Their Benefits – Part 3: Cost and Deployment

The final instalment of this trio of articles looks at the features of the two cloud server deployment models, public and private, as well as discussing how they can deliver real cost savings to their customers.

Cost Efficiencies

As mentioned previously, the responsive scalability of pooled cloud servers means that cloud services can offer significant cost efficiencies for the end user – the most salient of which is that the client need only pay for what they use. Without being bound by the fixed physical capacities of single servers, clients are not required to pay up front for capacity which they may not make use of, whether it be their initial outlay or subsequent steps up to cater for increases in demand. In addition, they avoid the set up costs which would otherwise be incurred by bringing individual servers online. Instead any set up costs generated when the underlying cloud servers were brought online are overheads for the cloud provider and are diluted by economies of scale before having any impact on their pricing model.? This is particularly the case as many cloud services minimise the effort and expense of specific cloud server and platform configurations by offering standardised services into which the client taps.

Lastly, cloud models allow providers to do away with long term lock-ins. Without the longer term overheads of bringing individual servers online for individual clients and maintaining them there isn’t the dependency on those clients for a return on that investment from the provider’s point of view.

Deployment

There are two common deployment models for cloud services which span the service level models (IaaS, PaaS, SaaS) described in part one:?Public Cloud:and Private Cloud.

Perhaps the most familiar to general population, and also the most likely to deliver some of the features and benefits mentioned previously, is the typical public cloud model. This model utilises the large number of pooled cloud servers located in data centers, to provide a service over the internet which members of the public can sign up for and access. However, the exact level of resource – and therefore capacity, scalability and redundancy – underpinning the each public cloud service will depend on each provider. The underlying infrastructure, including servers, will be shared across all of the service’s end users whilst the points at which the service can be accessed are open to anyone, anywhere, on any device as long as they have an internet connection. Consequently, one of the model’s key strengths, its accessibility, leads to its most prominent weakness, security.

Services which need to implement higher levels of security can instead use private cloud models. The architecture of private clouds can vary but they are defined by the fact that the cloud is ring-fenced for the use of one client. Servers can either be located in a data center, and accessed via leased lines or trusted provider networks, or on the client’s premises, and accessed by secure local network connections. They can be provisioned as either physical or virtual servers, but they’ll never be shared across multiple clients. Access to the servers and the cloud service will always be behind the client’s firewall to ensure that only trusted users can even attempt to use it.

Private clouds, therefore, offer greater levels of security (depending on the exact set up), but utilising smaller pools of servers means that they cannot always match the economies of scale, high capacities, redundancy and responsive scalability of public cloud models. Although, these qualities can still be achieved more readily than more traditional fixed capacity server configurations on local or trusted networks.