gLite: leading the development and adoption of standards

The Enabling Grids for E-sciencE project supports the high-throughput data-processing needs of the European research community and their international collaborators, supporting over 17,000 users across 160 projects and covering domains such as astrophysics, biomedicine, computational chemistry, earth sciences, high energy physics, finance, fusion, geophysics and multimedia. EGEE also supports business applications in the geophysics, finance and plastics industries.
The middleware that makes this possible is gLite, a software distribution produced by EGEE and composed primarily of open-source software from many sources – most developed within the project, some from external providers. gLite allows researchers to access physical resources attached to the EGEE infrastructure – including disks, computers, and instruments  – using their own applications. gLite services now have stable interfaces, often in the form of web services and defined by technical standards, which encourage application and tool developers to build their own software on top of the gLite platform.

The advantages of standards
Adoption of distributed computing infrastructures is expanding across all disciplines, making it increasingly important to have standard interfaces. Standards allow developers to protect their investment in the software they develop by enabling access to any infrastructure that supports these standards. Standards prevent developers from being locked-in to one particular platform. Standards encourage competition between middleware providers, allowing resource providers to replace under-performing software components with better components that use the same standard interface, with no disruption to the user community.
As EGEE’s infrastructure has matured, the project has increased its investment in standards and its liaison with standard organisations. EGEE uses many protocols common to the internet (e.g. TCP and IP), the web (e.g. http and https) and web services (e.g. XML, SOAP and WSDL); these protocols will not be covered in this article.
Instead, we will concentrate on the protocols that allow secure access to shared resources. These protocols come primarily from the Open Grid Forum (OGF), of which EGEE is an active technical participant, and the Organization for the Advancement of Structured Information Standards (OASIS) and Internet Engineering Task Force (IETF), where EGEE and the grid community are primarily external consumers.

Standards within EGEE: Security
The resources that make up the EGEE infrastructure are extremely valuable and access to these resources needs to be strictly controlled. EGEE, in common with other infrastructures, use the X.509 certificate and X.509 proxy certificates from the IETF to represent the permanent and delegated identity of an individual. Through the EGEE Virtual Organisation Membership Service (VOMS), the X.509 proxy certificate can include a SAML (Security Assertion Markup Language) assertion from OASIS, which provides information that other services can use to make authorization decisions. Authorization requests use XACML (eXtensible Access Control Markup Language) from OASIS to define the content of messages in the SAML message format passed to the authorization service.

Standards within EGEE: Information
With many thousands of services potentially available, discovering which one to use presents many challenges. The GLUE 2.0 specification from OGF provides a defined abstract information model that uses a set of known attributes with defined semantics to facilitate interoperation between grid infrastructures. These attributes can be rendered through a number of defined protocols. EGEE is currently adopting GLUE 2.0 with deployment planned for production use in early 2010.

Standards within EGEE: Data
Many researchers use EGEE’s infrastructure to analyse data stored in files. EGEE, in common with many other infrastructures, uses the GridFTP v2 protocol (an OGF specification that extends the FTP protocol) to provide high performance transfer of large data files over wide area networks. The storage resources at the source and destination of the file transfer are manipulated through OGF’s Storage Resource Management (SRM) protocol, which governs access methods when moving files to and from the permanent storage space (disk or tape) to the temporary space accessible to the GridFTP server.

Standards within EGEE: Computing
Nearly all communities supported by EGEE analyse file-based data by starting an application that processes that data. Two specifications developed within OGF and relevant to EGEE are the Basic Execution Service (OGSA-BES), which describes an interface for running and controlling applications, and the Job Submission Description Language (JSDL), which specifies how applications should be run. EGEE is examining how these specifications could be adapted to serve its users.

Standards within EGEE: Accounting
Since EGEE’s resources are contributed by different organisations for shared use by third parties, it is important that this use is accounted for. Within EGEE, and other infrastructures, OGF’s Usage Record specification is used as part of the protocol for transferring records between resources and the central accounting server. Partners within EGEE are examining the use of the Resource Usage Service as an interface for retrieving stored records, but this needs further development.

Summary
The adoption, development and implementation of standards is crucial to the health of the grid computing community. EGEE has contributed significantly to this work, thanks to its experience in running a large e-infrastructure. We are cooperating with other infrastructure providers within OGF’s Production Grid Infrastructure Working Group (PGI-WG) to ‘fine tune’ these specifications to our particular use cases and thus build interoperable e-Infrastructures.
The encapsulation of this ‘know-how’ into published use cases and the specifications of organisations such as OGF may represent a significant part of EGEE’s eventual legacy. This work also provides a route for new software providers to enter the grid world, and allows resource providers to deploy any software solution that meets the required standards.
Standards provide an essential foundation for allowing independent National Grid Initiatives to federate their resources into a European Grid Infrastructure, ensuring Europe’s researchers have access to the computing resources they need for many years to come.