Of all the IT industry topics, Cloud Computing is probably one of the most over-hyped and yet most under estimated in terms of its effect as a computing delivery model. Why is this?
Well - over-hyping is a fact of life in IT. Gartner Group's hype-cycle has long articulated the way that new developments in IT gain traction, attention from the media, and peak with inflated expectations. The subsequent decline into reality and then the slow rise into the plateau of business benefit realisation is well grounded in our real-life experience.
In recent conversations with leading Cloud Computing companies, we have explored the practical constraints to growth that may affect the traction of Cloud-based services. If, as the hype cycle suggests, there are some serious realities yet to address, then what are they? How can these be anticipated, considered and overcome?
Our point of view here is that there are 'limits to growth' in the adoption of Cloud Computing. Initial adoption of Cloud Services has been the low-hanging fruit of commodity style services such as e-mail, collaboration services, file sharing, etc. and non-mission critical services such as managed test.
As a utility service, Cloud Providers have offered many 'as a Service' consumption models such as - IaaS (Infrastructure), PaaS (Platform), SaaS (Software Apps) and BPaaS (Business Processes). What can be observed here is the relentless drive up the value chain - as the IaaS and PaaS offerings become increasingly commoditised, the survivors in the over-populated Cloud Services market will be those who concentrate on the value-add, differentiated services. It is these services that require careful consideration before transitioning core business systems and workloads into others hands - regardless of their brand prominence.
Limits that arise will increase as organisations start to consider core application workloads and data as Cloud candidates. It is these core services whose transition to the Cloud will hit limits. We offer the following thoughts and insights as, potentially, they all present ‘limits to growth’ that must be recognised and managed.
Firstly, end-to-end service management remains as an over-riding requirement. That is, to manage and offer a seamless set of business services regardless of sourcing and delivery. With every Cloud Service that is bought-in, no matter what its position in the IaaS/PaaS/SaaS stack - we need to recognise that as soon as the service has transitioned from internal to external, there is a new service boundary. Thus, the provider will have made contractual commitments for a price and some form of Service Level Agreement (SLA). If the Cloud Provider SLA is not aligned to the service delivered to the business, then problems can arise. For example, if there is a cloud service failure - and these are not uncommon - what commitment are we able to offer the business that the service will be recovered in a particular time; or that crucial business information will be recovered - and by when?
The way to address the Service Boundary limiter is to know that it's there and plan how this will be identified and integrated into the overall service. It may be that the cost savings of a Cloud Service may be so compelling that the business is prepared to compromise service levels. If this is the case then a clear communication and business engagement process must be undertaken to ensure that this - and its implications – are well understood.
We should also recognise that we may move a service over time from a Public to a Private Cloud or from a Hybrid to a Private Cloud, etc., as we gain confidence, understand risk and as Cloud Providers mature and extend their capability. Understanding the Service Boundary for the first stage will then re-apply to the next stage to allow a more structured reassessment of the next part of the cloud journey.
It is also vital to consider where your business data will reside and how it can be accessed, integrated and recovered if necessary. There are some significant business decisions that need to be in place before relinquishing your data into the care of others. It is, after all, the very core asset that allows you to know your customers, the position of your business, and your strategy for the future.
Finally, there is no certainty that any one Cloud Service Provider will be around over the long-term – especially with the current crowded marketplace at. Undoubtedly, there will be mergers, acquisitions and other market exits – all of which could disturb the service boundary, or its commercial terms. The key to successful use of cloud-based services is having a stable long term interface - if I'm going to bet part of my business on this service I need some sort of guarantee that it's still going to be around in 10 years time, so I can build it into my wider IT ecosystem.
In summary, although it seems that the Cloud is now a daily reality, this could still be a perception and an effect of over-inflated expectations - a la Hype Cycle. If there are limits to growth in Cloud Service adoption, then our point of view is that this will be manifested in service integration and in the conscious choice of where to place and manage business data and the service boundary. If this is recognised and addressed as part of the evaluation process, then perhaps we avoid the 'trough of disillusionment' and venture with more confidence onto the 'plateau of enlightenment'!