Software defined infrastructures : implications for technology and business strategies for competing in the era of hyper-scale computing

The adoption of Cloud Computing and the emergence of new Cloud services pioneered by Amazon (AWS) have brought the importance of agility and flexibility of infrastructure to the forefront. Companies ranging from small to large Enterprises today have a Cloud Strategy. Their Cloud Strategy ranges from being all-in with respect to moving their internal IT infrastructure to the Cloud to moving only specific low SLA workloads to the cloud. However, not everyone can or is comfortable letting their sensitive data leave their infrastructure and reside on third party infrastructure that is not in their control. This has led to customers building Private Clouds, which however don't give them the scale or the flexibility that Public Clouds provide. So customers are now looking at ways to replicate the success of Public Clouds in their own environments. What they find is that the existing IT infrastructure and its architecture is inadequate to provide those benefits. To achieve the Public Cloud characteristics, customers have started looking at the infrastructure built by Hyperscalers like Google, Amazon, Microsoft, Facebook and the benefits that they have been able to achieve as a model to build their own IT data centers. The infrastructure of these Web giants consists of commodity hardware components managed and driven by intelligent Software. This has given rise to various Software Defined technologies like Software Defined Networking and Software Defined Storage. As the customer interest and adoption of these technologies increase it presents a huge business challenge to existing IT equipment vendors. Not only are they faced with technological challenges as the architecture moves in a different direction than they were charting to but it also presents a business model change which if not navigated carefully can lead to significant erosion of their revenues. This thesis identifies the impact of Software Defined Infrastructures on the enterprise equipment vendors and proposes strategies for successfully competing in the age of Hyper-scale Computing. Thesis Supervisor: Patrick Hale Title: Director, System Design and Management Fellows Program. Senior Lecturer, Engineering Systems Division