We use virtualization to study and develop the foundation technologies to build large-scale distributed computing infrastructures capable of hosting thousands of applications that deliver IT functions to millions of users. Building such infrastructures requires solving the following key challenges: scalability, complexity and flexibility. Scalability refers to the ability of harnessing computing powers spread across a large number of distributed resources. Complexity refers to the problem of managing the configuration, security, availability, and performance of a large number of distributed applications, and systems. Flexibility and speed of delivery refers to the ability of creating and deploying new IT services at a speed one order of magnitude faster than what traditional IT infrastructure achieve today.
We are leveraging virtualization in the following ways:
- Image Management. Virtual images, i.e., pre-built software stacks, will become the new unit of distribution, deployment, licensing, maintenance, archival and service/support. Virtual images will change the way we compose and configure software solutions.
- Goal oriented resource optimization. Virtualization will extend beyond single systems to multi-system pools consisting of servers, network and storage, thus creating a new platform for integrated management and optimization of data center resources.
- Model based solution composition and management. A new breed of tools will emerge to allow datacenter administrators to quickly assemble solutions from ready-made virtual images and pre-built templates. These tools will allow best practices experts to define deployment patterns that the tools can use to automatically generate actionable distributed deployment models based on administrators inputs and using model transformation techniques.