Scaling Cloud Structures

Keith Rochford - Dell

Cloud is no longer the nebulous Next Generation Delivery model we have all become accustomed to talking about. Cloud is upon us. It is here and now, and every day we use and consume Cloud technologies and practices to make our lives better. As a result we open up to new ways of consuming and mixing cloud in order to gain maximum benefit. Private cloud has allowed organizations to meet regulatory compliance, governance, and security requirements while affording the agility and manageability expected in today’s responsive environments. With Private and Hybrid Cloud, we can drive consolidation across campus, regions and even global platforms freeing up data center space and offering a consistent experience for consumers.

Intelligent application of Cloud resources allows us to achieve new efficiency in day to day, academic and research computing. By applying the correct workloads at the correct time onto the correct stamp, we can increase efficiency, agility and flexibility with the control and governance required to help drive IT away from the traditional rigid cost center model of the past, and toward the more innovative, responsive and cost efficient service centric model of the future.

When we look to scale and deploy workloads across Cloud platforms such as Private and Public Cloud we need to determine which applications are the best fit for each platform based on requirements such as security, latency and application dependencies. Applications written 15-20 years ago may not be able to leverage the benefits of virtualisation without significant re-factoring. Further, even if we do look at application optimization for the Virtual infrastructure of today, this does not necessarily mean that because it will port and scale within the parameters of our “Private Cloud” infrastructure, that it will do so as seamlessly, or indeed at all, across the scalable and elastic expanse of today's “Public Cloud” platforms.

We need to be mindful when selecting “Hybrid” Cloud as a delivery model for our application sets. Critical to the success of such a delivery model is the understanding and classification of our applications and their fundamental dependencies and how the “Hybrid Cloud” will either exploit or spoil the performance potential of the application. The fundamental prerequisite to a scalable application or service is a scalable and responsive underlying architecture. A scalable architecture is one whereby the various components of the architecture are independent of each other allowing the development of loosely coupled, distributed applications capable of scaling to extraordinary levels.
Deploying Cloud platforms and using Public and Private delivery models demands intelligent and accurate design principles and must seek to leverage existing skillsets, infrastructure and processes to compliment and enhance IT. As with all production infrastructure security is of paramount concern. By understanding the applications and being exact in our planning and execution, identity management access control and governance, we can ensure the level of confidence required to embrace the full potential of this powerful delivery model.