Linked Enterprise Data Patterns
Data-driven Applications on the Web Dec 6-7 2011, Cambridge, MA, USA

Call For Participation

Abstract

The four rules of Linked Data provide a very simple guide for publishing data on the World Wide Web. This led to a remarkable evolution in the domain of publicly available data, commonly known as "Linked Open Data". Applications in that space are beginning to emerge, usually relying on the public availability of various Linked Open Datasets.

Linked Data technology also offers a huge potential for enterprise applications, e.g., for the integration and the management of data within and across enterprises. The distributed nature of Linked Data enables loose-coupling for data sharing within and between organizations. With Linked Data, enterprises have a unique opportunity to cooperate in their use of shared data without the costs of extensive coordination.

Sharing a common data model (RDF) allows us to establish design patterns for providing dereferencable resource identifiers, migrating and cloning data as business needs and data authorities evolve. These will address issues such as data distribution, query federation, access control, encryption and signature, legal problems around the access of datasets and business models in using open or closed linked data. Please join the W3C Linked Data community at this workshop to air requirements, share solutions and develop a healthy and scalable Linked Enterprise Data infrastructure.

Note that the scope of this workshop differs from that of the W3C Workshop on Data and Services Integration. This workshop focuses on design patterns for Linked RDF Data in the enterprise.

Background

Integration of heterogeneous data and services has always been a concern for applications and enterprises. The diversity and opportunities have grown exponentially along with the web.

The standardization of RDF provides a powerful foundation for the globalization of data. The distributed extensibility model reduces the costs of both standardization and extension. The web itself provides a robust architecture for creating, transferring and updating RDF graphs. The RDF query stack, including SPARQL 1.1 Query Language, SPARQL 1.1 Protocol for RDF, SPARQL 1.1 Update and SPARQL 1.1 Graph Store HTTP Protocol, enable certain usage patterns, but common application infrastructure models like partial update, bulk update publish/subscribe still need development.

Many applications are shifting towards data-centric protocols, enabling a loose coupling between applications using different aspects of shared data. The schema languages RDF Schema and OWL help application developers work within a contract for a shape of the data. Business process management must address not only static contracts, but the evolution of contracts.

Scenarios

Tool integration has historically been addressed via individual bridges between pairs of tools, requiring development and maintenance of an exponentially increasing number of shims. Overhauling the traditional pair-wise integration approach to application integration requires that we define, on top of the current Linked Data standards, a the set of conventions that are of potential use to a much larger audience.

Linked enterprise data infrastructure must address existing intra-application communication needs as well as ones emerging from the newly available deployment of Semantic Web tools and know-how:

  1. Transfer of large resources: A resource is sometimes too large to be reasonably transmitted in a single HTTP message. Chunked serialization formats and paging mechanisms can allow clients to tractably retrieve representations of large resources.
  2. Resource containers: It is frequently practical to create containers grouping related resources with one another. Independently of the type of resources being managed, containers invariably need to provide a set of methods to retrieve the list of existing resources they contain, to add a resource, etc.
  3. Authority evolution: As Linked Data becomes more critical to business processes, custodanship must migrate without breaking deployed infrastructure. Solving this will enable organizations to develop limited partnerships to share development.
  4. Incremental service deployment: Many organizations deploy applications to a subset of the users before unleashing it upon the total user base. Robust Linked Enterprise Data will allow the useful dereference of resources as their "home" (e.g. URL) potentially moves between machines and custodians.

Goals and Scope

The goal of this workshop is threefold:

  1. Describe a set of standard patterns, design choices and best practices that will give application writers much stronger guidance and reduce the number of design decisions they have to make when adopting Linked Data as an application architecture.
  2. Identify a handful of gaps in the current Linked Data specs that are uncovered by common application scenarios.
  3. Propose a set of simple solutions to those gaps that could be rapidly specified and approved by W3C.

Topics for position papers may include, but are not limited to:

The workshop will discuss the similarities, gaps and differences between the different Semantic Web approaches. (We will not focus on ontology development or inferencing as they would likely derail the "plumbing" discussion.) Participants will, in particular, include practitioners, technologists, and architects. Presentations are expected to be technical in nature; those interested in persuing business cases may wish to engage in the W3C business or community group process.

The outcome of this workshop will be reported to the current related working groups, RDF-WG, Provenance WG, GLD WG, and may be used as input for chartering other work.

Participation Requirements

All participants are required to submit a position paper by 31 October 2011. W3C membership is not required to participate in this workshop.

The total number of participants will be limited. To ensure diversity, a limit might be imposed on the maximum number of participants per organization.

Instructions for how to register will be sent to submitters of accepted position papers. These instructions will also indicate a possible limit on the maximum number of participants per organization.

Workshop sessions and documents will be in English. Position papers, presentations, minutes and the workshop report will be public.

There is no fee to participate.

Expression of Interest

To help the organizers plan the workshop: If you wish to participate, please as soon as possible send a message to team-workshop-submissions@w3.org with a short (one paragraph) "expression of interest" stating:

Note: Sending that expression of interest does not mean that you registered for the workshop. It is still necessary to send a position paper (see below), which then must be considered for acceptance by the Program Committee.

Position Papers

You paper must meet the following criteria:

Based on a review of all submitted position papers, the Program Committee will select the most relevant and invite the submitters of those papers to the Workshop. From among all accepted papers, the program committee will choose a small number of papers judged most appropriate for fostering discussion, and ask the authors of those papers to give short presentations about them at the Workshop. After the workshop, those presentations will then be published on the workshop home page.

Important dates

Date Event
31 October 2011 Deadline for position papers
7 November 2011 Acceptance notification sent
14 November 2011 Program released
noon (EST) 2 December 2011 Deadline (for lunches and name tags) for Registration
6-7 December 2011 Workshop

Workshop Organization

Workshop sessions and documents will be in English.

Chairs

Program Committee

Venue

The Workshop will be at MIT's Stata Center in Cambridge, MA, USA. More detailed venue information will be made available closer to the event.

Deliverables

Position papers, schedule, accepted presentations, and report will also be published online.