Background

OMII-UK was formed in January 2006 as a partnership between the existing OMII activity at the University of Southampton, and software activities at the University of Manchester within the myGrid project and the OGSA-DAI project at the National e-Science Centre and EPCC. It has a budget of around £2.5M for commissioning the development, hardening and integration of software to support its various user communities. We are now seeking proposals for work:

• in the area of software infrastructures and tools to bring campus, national and international grids to the desktop through OMII-UK services,
• in the area of portlet development and integration with OMII-UK services,
• to support the development and standardisation of Grid APIs.

One of OMII-UK's roles is to commission software components (through its 'Commissioned Software Programme') from the community to meet community needs - as such we may be only willing to fund certain work items from a proposal. We are not able to fund general research activity, nor are we likely to fund the development of new prototypes. We are interested in funding the hardening of advanced research prototypes that have already demonstrated uptake within the e-Science community so as to accelerate their adoption by a wider user base. We generally expect the work that we fund to be open-source and to be accessible to other developers within the community - both during and after our funding.

Submissions - 27/11/2006

Proposals MUST follow the structure described on the website and in the associated documents (http://www.omii.ac.uk/projects/funding.jsp). We strongly encourage you to provide a letter of intent describing the proposed work and how it would be used in your, and others projects. This will allow us to provide early feedback on the proposed work and ensure that any subsequent proposal targets areas that are of interest to us. These proposals will be reviewed internally by OMII-UK for technical feasibility and user engagement. Typically, projects will be 12 months in length and might involve 24 person months of effort. Longer and larger projects may be funded if justified but this will be the exception rather than the norm and a strong case will have to be made. The closing date for this call is 07/01/07 with funding decisions being made by 05/02/07 and projects starting April/May 2007. Informal enquires and 'letters of intent' about the work may be made to the Director (Steven Newhouse) director@omii.ac.uk. Completed proposals should be sent to Mrs Gill Schofield (gs2@ecs.soton.ac.uk). Applications may be made by non-UK based development groups but substantial engagement and benefit to the UK community will have to be made. It is expected that these groups will be experienced in producing high-quality software using open-source development methodologies.

Bringing Desktop Access to Grids

Providing access to remote compute and data resources is a fundamental issue that is now facing today's scientist and researcher. Portal interfaces (e.g. Virtual Research Environments) do not always provide the flexibility for advanced use, e.g. advanced data search and analysis options. Instead, interfaces which can provide such flexibility are needed in order to compose resources into workflows that can be orchestrated directly from the desktop. In many areas such interfaces already exist (or are under development) and these are not the focus of this call. Seamlessly accessing this emerging e-infrastructure, whether located on the campus, as a national service (e.g. the UK's National Grid Service), or as part of international collaborations (e.g. EGEE or TeraGrid), is becoming a key user requirement. Such access should be provided from within an environment that the end-user is already comfortable (e.g. Bash scripts, Matlab, R, Python, Perl, Java, C, GUIs, etc) and to leverage APIs, such as SAGA (Simple APIs for Grid Applications), where relevant.

OMII-UK is seeking to support the integration of off-desktop resources into such environments through desktop environments (as opposed to a portal-based virtual research environment) commonly used by scientists and researchers. We are interested in supporting projects that address one or more of:

• Accounting: As users start making regular use of services from different providers it is essential that such use is recorded, aggregated and returned to the interested parties respecting their privacy, e.g. the support of anonymised usage aggregation through the Resource Usage Service (RUS) between campus and national service providers. The focus should be on bridging existing systems and making the record of any use available to that individual.
• Computation Integration: Including connecting GridSAM to distributed resource management (DRM) infrastructures that it is not currently connected to that are deployed within campus and national grids, e.g. PBS, GridMP, gLite.
• File Integration: Providing transparent access to and movement of stored files regardless of their location. For instance, input data files stored on the desktop (or an accessible file store) should be securely accessible from remote resources, and the output from these computations should either be placed or retrievable back to the user's desktop with few explicit operations by users or their applications. This may require supporting work to capture meta-data relating to the generation of the data file to support curation activities. Data movement may be 'pushed' from the desktop to the resource, or 'pulled' by the resource from the desktop.
• Packaging the job submission and data management into integrated operations that suit certain computation patterns, such as parameter sweeps.
• Resource Selection Across Grids: Tools and services that provide transparent resource selection and execution to the end-users across different grid infrastructures by integrating existing brokering solutions. Work may be required to extend these existing brokering solutions to make use of emerging standards such as the Basic Execution Service (BES) and Job Submission Description Language (JSDL) from the Open Grid Forum.

We expect proposals to be driven by use cases that will provide substantial benefit to an identified user community. There are many related technology solutions that may provide partial implementations to some of these problems. We would expect these to be used where appropriate with any extensions contributed back to the community, and the eventual integration to be contributed into the OMII-UK repository, and potentially the OMII-UK software release. The authentication and authorisation technologies used in the proposed work, should where ever possible, use infrastructures already in use on campus or national resources, e.g. X.509 certificates or Shibboleth.

Submissions - 14/11/2006

Proposals MUST follow the structure described on the website and in the associated documents (http://www.omii.ac.uk/projects/funding.jsp). We strongly encourage you to provide a letter of intent describing the proposed work and how it would be used in your, and others projects. This will allow us to provide early feedback on the proposed work and ensure that any subsequent proposal targets areas that are of interest to us. These proposals will be reviewed internally by OMII-UK for technical feasibility and user engagement. Typically, projects will be 12 months in length and might involve 24 person months of effort. Longer and larger projects may be funded if justified but this will be the exception rather than the norm and a strong case will have to be made. The closing date for this call is 10/12/06 with funding decisions being made by 12/1/07 and projects starting in March/April 2007. Informal enquires and 'letters of intent' about the work may be made to the Director (Steven Newhouse) director@omii.ac.uk. Completed proposals should be sent to Mrs Gill Schofield (gs2@ecs.soton.ac.uk). Applications may be made by non-UK based development groups but substantial engagement and benefit to the UK community will have to be made. It is expected that these groups will be experienced in producing high-quality software using open-source development methodologies.

The current calls for proposals are:

1) Portal Support within the OMII-UK software release

OMII-UK provides a collection of software components and tools that have been integrated together, tested and documented. We would like to provide support for web based interaction with these software components. We will provide an integrated JSR-168 portlet hosting environment in the Tomcat environment shipped within the OMII-UK software release upon which any funded portlet work would be based. We would expect this work to possibly include:

• the customisation and/or extension of existing mature JSR-168 portlets,
• the use of WSRP (Web Services for Remote Portals) to expose portlets for remote access,
• the use of WSRP to federate content from other portlets or providers.

All work must be integrated to work within the OMII-UK software environment and be demonstrated to provide access to OMII-UK services where relevant. We are interested in supporting portlets relating to our core activity:

• Job Submission and Job Management: Using the Basic Execution Service and JSDL specifications to submit jobs into GridSAM, or any other compliant service.
• Service Browser: Presenting information retrieved from a UDDI compliant registry, e.g. Grimoires.
• File Transfer: Movement of files between different locations. This could include interfaces to both SRB (Storage Resource Broker) and/or GridFTP based services (e.g. RFT - GT4 or FTS - gLite).
• Workflows: Portlets to manage the contribution, editing, searching, download, execution, monitoring and community annotation of workflows.
• Data Access: Use of WS-DAI,WS-DAIR, WS-DAIX and supporting specifications to interact with and manage OGSA-DAI data services.

In addition we are interested in considering support for portlets providing:

• Account Registration: Integration of portlets that handle user registration (e.g. GAMA, PURSe, etc.) with the 'Account' service in the OMII-UK software release.
• Attribute Management: A common authorisation mechanism within Virtual Organisations is to use attributes to drive authorisation decisions.
• Policy Management: Define policies to specify access to services/portlets using a recognised policy language (e.g. SAML, XACML, etc.).
• Certificate Management: Management of certificates stored within a MyProxy server.
• Collaboration Environment: Portlets that support interaction between distributed groups, e.g. wikis, message boards, etc.
• Execution of Legacy Applications: Provision of a portlet to specify command line arguments to legacy applications (using JSDL) and a portlet to manage the definition of the meta-data to specify the command line structure.

2) Grid Programming Interfaces

One of the current barriers to broader adoption of the Grid paradigm amongst application developer communities is the continued evolution and variation of 'client-side' programming APIs between different middleware distributions. Early work within the Open Grid Forum (OGF) in this area included the Distributed Resource Management Applications API (DRMAA) and GridRPC working groups. Activity in this area has been extended to define a middleware agnostic programming interface to the core capability offered by different middleware providers. Work in this area is co-ordinated through the Simple API for Grid Applications Research Group (SAGA-RG) and a detailed API specification is being developed by the SAGA-Core-WG. More information can be found at:

http://forge.gridforum.org/sf/projects/saga-rg

http://forge.gridforum.org/sf/projects/saga-core-wg

We are seeking proposals to support engagement with this standards activity by providing reference implementations of the SAGA, DRMAA and GridRPC APIs to several Grid middleware distributions. Groups would be expected to provide reference implementations of the SAGA Java API to the OMII-UK software release and one or more other middleware distributions, (e.g. OMII-UK, the Globus Toolkit, gLite, CROWNGrid, etc.) so as to demonstrate interoperability between different middleware, and to actively contribute their experiences back to the OGF groups to further refine and improve the specifications.

In addition, proposals covering other programming environments (e.g. PERL, Python, C#, etc.) would be welcomed providing sufficient user engagement and demand can be demonstrated. Work contributing to the SAGA activity that standardises APIs beyond those currently being worked on by SAGA (e.g. steering, data access, etc.) would also be considered. Groups proposing new APIs would be expected to show motivating use cases across several communities and to have had experience in standards development activity. The APIs that would be considered for funding in this area will need to demonstrate some community convergence and have maturing implementations - support for research activity will not be considered.