Cloud Interoperability Testbed

Project Information

Discipline

Computer Science (401) 

Orientation

Industry 

The Cloud Interoperability Testbed will serve as a mechanism to host interoperability tests for different machine control, data transfer, resource reporting and usage agreement standards and implementations of other new standards efforts. The primary purpose will be to give developers an opportunity to try out implementations of code that implements either server or client functionality for the use of multiple standards from different standards development organizations. The initial effort for this project will be focused on implementations of the Open Cloud Computing Interface (OCCI) from the Open Grid Forum (OGF), Cloud Data Management Interface (CDMI) from the Storage Network Industry Association (SNIA), the Cloud Infrastructure Management Interface (CIMI) and the Open Virtualization Format (OVF) from the Distributed Management Task Force (DMTF). Other standards will be added as the project proceeds.

A large amount of effort is being expended by multiple organizations to develop standards in the area of distributed computing , spanning many specific topics ranging from advanced networking control to infrastructure management to data transfer and packaging protocols. In the absence of a common framework in which these can be tested, development activities for implementations of such standards are necessarily fragmented and limited to those chosen by a particular development team. The use of a virtualized space made available through FutureGrid provides a range of opportunities to streamline tests and coordinate activities to promote common use and interoperability of code written to such standards. This project will complement the activities of the Open Grid Forum Cloud Interoperability Working Group (CI-WG) and the OGF/SNIA Cloud Plug-Fest working group, the latter of which also includes participants from members of DMTF.

By providing a forum and project explicitly oriented towards early testing of standards and code that implements them, we provide a vehicle for involvement of students and other participants who do not yet have experience in the tools and methods of advanced distributed computing. A significant amount of the code and material that will be posted as a result of this project will be oriented toward beginners and those who would like to participate in modern cloud code development but who do not yet have access to their own server resources. This will significantly increase the ability of students, researchers and startup businesses to access advanced implementations of modern current standards and code and to see and test implementations of these standards directly. Once the basic concept is in place and working, we expect to proceed forward with additional projects that build on this effort are explicitly oriented towards outreach and education.

Project Contact

Project Lead

Alan Sill (alansill) 

Project Manager

Alan Sill (alansill) 

Project Members

Gary Mazzaferro, Michael Kretzschmar, Alexander Papaspyrou, Florian Feldhaus, Piotr Kasprzak, Yugendra Guvvala, Ilja Livenson, Anupam Tambi, Andy Edmonds, Christian Karnath, Rajendar K, Keyun Ruan, Akash Pargat, Eugene Luster, Iain James Marshall, Mark Carlson, Soheil Mazaheri, Chaz George  

Resource Requirements

Hardware System

• Not sure

 

Primary ability needed is to be able to spin up test machines that can be provisioned as needed with software stacks implementing various standards. Long-term running of services is likely not needed. We expect to be working with members of the OpenNebula, OpenStack, XSEDE, Globus and Nimbus software development teams, among others, and so may need to deploy test instances of future versions of these software stacks.

10-20 developers anticipated in the early stages. Use may grow depending on the success of the project, but we intend to caution developers that these are expected to be test implementations and not long-term development servers at the moment. It is important, however, that access to the project remain possible for an extended duration (1 year or more preferred) even though individual machine instances for any particular implementation used within the project are expected to run only for a limited amount of time.

Project Timeline

Submitted

11/29/2011 - 12:16