Cloud computing = do a lot more computing with a lot less money. But it takes a clear direction and a lot of work to get there. Instead of uniquely deploying technology solutions for each product or investment, standards that take advantage of ubiquitous computing need to be rigorously deployed.
In my post about The business potential in the ‘nexus’ enabled market I started looking at the business impact of cloud computing. Let’s delve a bit deeper into this, and I’ll also add some of the technology advances and challenges that I see.
My definition of cloud computing is the ubiquitous availability of computing resources, operating in an environment that is rapidly available, scalable, secure and reliable. Cloud computing is available internal to companies or external through service providers, and can be a private or public infrastructure. Most companies will either need or have a hybrid solution to address different needs.
The cost of scalable computing – Both the cost and the scalability are of great importance. Forget about the technology itself for a second, accept that processing of transactions and other data can happen real-time or near real-time in a different expense paradigm.
- There will be a combination of reduced infrastructure expense and increased availability of computing, depending on needs. To achieve this, the infrastructure architecture needs to take a practical yet rigorous approach to applying standards so as to fit the need into the cloud computing capabilities. In my experience, at least 80% of computing needs can be accommodated with cloud computing.
- Computing requirements will gradually become more standardized at the technology level and rely on ubiquitous computing availability. Technologists love to divide into layers, for example the network layers, the infrastructure layer, the application layer and the data layer. A decade or more ago the network layer was difficult and expensive, now it is much more ubiquitous and less expensive. The same will now happen with the infrastructure or computing layer through the use of standards and cloud computing.
- This will also enhance reliability and disaster recovery capabilities both of which are still significant challenges in the industry. Once standards are applied and in operation, it will be a lot easier to deploy, support and allow for redundant operations for reliability and disaster recovery. For one system I deployed across the world a number of years ago with one in Asia, one in Europe and two in the US it became a nearly impossible and 12+ months task to set up, in coordination with the other locations and each with disaster recovery locations. Today with cloud computing this can be done once, and deployed easily.
Real time analytic engine – Analytic engines require significant compute power and rapid data systems, both of which become significantly more available in the cloud paradigm.
- New predictive analytic technology combined with the compute availability will shift capabilities from historical limited analytic systems that run on a periodic basis to real-time or near real-time systems that have greatly enhanced data association and visualization strengths.
- The business impact of this is huge – in many current situations analytic solutions are not being applied to solutions as the combination of cost, data availability and timeliness do not allow for this.
One-time and ongoing operations – What we see today are the early applications in big data and analytic capabilities, and they mostly focus on the one-time analyses or data mining opportunities. This will evolve into operational capabilities that perform these functions on an ongoing basis. If previous experiences apply, it will take years to get to the point where technology and systems are available to deploy for ongoing operational capabilities.
Current Challenges – The core technology of cloud computing is reasonably mature, the most recent advances to operate a cloud within a larger infrastructure have helped easy availability. Current challenges are mostly focused around the administration and hybrid operation of the environments. There are still many administrative and operational steps needed to operate a hybrid cloud in any company environment. The other challenge is the architecture of how applications are ready to utilize the cloud infrastructure. Applications need to be designed or redesigned to take advantage of cloud computing.
Advances – The shorter term advances will be a direct result of the current challenges, thus enabling administration of the hybrid cloud in an easy and more effective manner, as well as driving architecture standards. The longer term advance I see relates to the enabling of data as a ubiquitous resource. Although our current and rapidly growing data availability is over 80% unstructured, we are still largely stuck in the structured, relational database technologies. Both unstructured and structured data will become available in the cloud infrastructure such that high volumes, current and comprehensively linked data is easily utilized.