There is an increasing requirement across research domains (Biomed, Physics, Social Sciences, etc.) to access large scale shared data sets and to provide access to collaborators to UoE data sets. For data set volumes of 10s of Terabytes to Petabytes (becoming increasingly common), special transfer services that can operate over long periods unsupervised, can make optimal use of available bandwidth and can recover from failures are required.   This investment will capture detailed requirements for individual locations, development of detailed computing architectural design. EPCC has purchased a UoE site license for Globus software on behalf of EIDF, which has been made available to ISG to provide service to other parts of the University.

The proposed project will define a service model for a Globus service, for example presentation of DataStore data sets and also local College presentations of Globus. Given the current security threat environment there are some considerations on the methods of presentation of the Globus services that need to be considered (Globus shares end link connections using their own AAI system based on end link declarations). A single open presentation for all of DataStore would not be acceptable from a data security point of view.

The model is expected to be based on the creation of specific VMs (on request) with specified access and time limited, the ability to create these VMs may be delegated to support staff in the Colleges to allow flexibility and responsiveness. However, these are the details that need to be addressed in the service definition and creation portions of this project. Similarly, if additional license costs (beyond the standard license already purchased) are required to users who demand higher levels of security then mechanisms for allocation of costs and access to these services will need to be defined.

This project will deliver this service. Additional project documentation can be found in the project SharePoint area.