Abstract

Cloud computing is often compared to the power utility model as part of a trend towards the commoditization of computing resources. However, today’s cloud providers do not simply supply raw computing resources as a commodity, but also act as distributors, creating cloud services that are not compatible across providers. In this talk, I will discuss a new cloud service distribution layer, called a Supercloud, that is completely decoupled from the cloud provider. A Supercloud gives its users the illusion of their own homogenized private cloud (albeit, layered on top of one or more third-party providers). Under the hood, the Supercloud can include different hypervisors, hardware architectures, storage subsystems, and connectivity fabrics. Leveraging a nested paravirtualization layer called the Xen-Blanket, the Supercloud maintains the control necessary to implement hypervisor-level services and management. Currently, the Supercloud has been deployed using resources from several major cloud providers, including Amazon EC2 , Rackspace, HP Cloud, and some private clouds. VMs run in a virtual network and can be migrated seamlessly across different clouds, with different hypervisors and device models. Using case studies we demonstrate that, being able to deploy applications to more regions and granting more control to end-users, the Supercloud can reduce latency and cost compared to the underlying cloud providers.