The attraction of cloud services is likely to increase amid continuing cuts to budgets in both the public and private sectors. Outsourcing to cloud-based providers (or ‘cloud sourcing’) allows companies to reduce the overheads associated with hosting services in house. But finding a successful model for this transformation takes planning, and although ‘the cloud’ has become a homogenous term for remote services, organisations do not have to adopt a one-size-fits-all approach.
Cloud is expected to be a major growth area in ICT, according to Analysys Mason estimates, with the worldwide market for enterprise cloud-based services forecast to grow by 75% to USD31.9 billion in 2017. There are benefits from off-loading ICT operations to a third-party in terms of reduced overheads and re-focusing on core business, but there are also some risks.
When assessing these risks corporate users must consider the importance of different sites within their estate and the relative business impact of individual sites losing access to corporate data and applications. For example, an organisation might be ready to accept the risk of loss of ICT services at some smaller branch offices, while regarding any loss of service at an HQ installation as unacceptable.
Listed below, and illustrated in the following figure, is the simplified example of how a global enterprise might use cloud across three categories of site (the categories could equally apply to individual organisations of equivalent sizes).
- Branch offices (small) – with no local servers, offering a simple ICT environment which is cheap and easy to support, but having total reliance on the WAN to assure connectivity to cloud services. We can call this category a ‘full cloud’ site.
- Regional centres (medium) – equipped with some servers to provide access to mission-critical applications and secure data, assuring business continuity in the event of a WAN failure (and consequent loss of access to the cloud). Access to less important applications and their data can be hosted in the cloud. This solution is sometimes known as ‘hybrid-cloud’.
- Global HQ (large) – standalone sites having no reliance on the WAN for access to applications and their data. Such sites would be immune from loss of access to the cloud.
Cloud services are only as reliable as the networks over which they are provided, and although WAN access circuits can provide ‘five nines’ availability (99.999%), there is no such thing as 100% availability. Therefore, organisations must perform a cost-benefit analysis, weighing up the provision of resilience against the risk to the business of a loss of service at each site. WAN access bandwidths must also be sized appropriately to ensure the associated network is able to support an application’s latency and data transfer requirements.
Figure 1: Categories of site and example levels of cloud adoption [Source: Analysys Mason, 2014]
Whatever model is used, our experience of supporting ICT migrations has shown that several key issues have to be addressed if the transition is to be successful. Organisations should ensure that:
- service contracts are well structured and based on recognised standards (including measurable SLAs and exit conditions)
- ownership of the assets is clearly documented and that equipment refresh programmes are defined
- communications links between the service provider’s data centres are secure, resilient and of adequate capacity
- the technical design has adequate input from the client, and that there continues to be in-house staff to monitor the contract.
These well-established rules should apply equally if organisations are considering infrastructure as a service (IaaS) or platform as a service (PaaS) from a cloud provider.
In short, corporate users can choose how much they depend on outsourced services (based on their size and criticality), and what level of service they want from a cloud-based provider. In fact, some form of cloud-based service can be available to everyone – if the business case stacks up.