Mikelangelo will disrupt cloud computing across the whole virtual infrastructure stack. This stack covers virtualization technology, operating systems, cloud middleware, big data stacks, and high performance computing (HPC). We will work to improve the I/O performance of virtualised infrastructures and applications running on those infrastructures. More concretely, we will improve and extend a hypervisor (sKVM), a new operating system (OSv), implement new communication methods via remote direct memory access (RDMA), and integrate with a cloud middleware and HPC batch systems. Thus, the project covers the whole software range of the modern computing stack for a broad set of use cases. These use cases span the applications in the fields of big data, HPC, and cloud computing.
We envisage significant improvements to the efficiency, security, and usability of cloud computing as a result of Mikelangelo. In practice, cloud computing relies heavily on virtualisation. Virtualisation offers near-zero overhead for computation. However, virtual machines (VM) have an efficiency of only about 60-70% for I/O operations. This overhead limits the applications that clouds can host reasonably. For example big data, HPC, and real-time applications are traditional domains in which virtualisation has found only limited application.
We will increase I/O efficiency with an improved hypervisor called sKVM. Furthermore, we will continue development on a new operating system, called OSv, which Cloudius Systems, one of the project members, build specifically for cloud computing. Both sKVM and OSv will receive further extensions that will allow efficient communication via RDMA. We will then integrate both sKVM and OSv with a cloud middleware, to provide the advancements in a productive environment to users of infrastructure services. This integration with the cloud middleware will include a novel application deployment model, based on OSv. Four use cases will leverage those advancements. One use case will offer big data clusters on demand. Another use case will offer cloud bursting. The remaining two use cases will offer HPC with VMs.
The project’s scope and its merits can be best explained in a bottom-up fashion according to its architecture. We are going to describe Mikelangelo’s architecture briefly in the following paragraphs.