IDM 2000 WORKSHOP REPORT

 

IR and DM in the WWW and Wireless Era

 

 

 

 

 

MISSION

 

                The advent of wireless and WWW access to data will dramatically impact the way in which databases of the future are accessed and managed.  Traditional database approaches that use static schemas and standard SQL query interfaces will not function in the future database environment.   Research into the issues associated with these two new database access paradigms is imperative if the true potential of both is to be met.

 

ISSUES

 

                There are many issues associated with the WWW and Wireless access of databases.  We briefly examine some of them below.

 

Wireless Issues:

1.        Wireless access to data in the future will require the ability to access the data totally based on the context.  Not only will there be drastic differences between the hardware and network resources available to the different database users,  but the context is variable and may be scarce.  The context may change in the middle of a query execution.  The key point is that the context is variable not just heterogeneous as in a distributed environment.

2.        Database queries  may have results dependent on the context.  Of particular interest are Location Dependent Queries (LDQ) whose result depend on location.  Particular research interests include: location binding, location leveling, indexing and directory structure.

3.        In Ad Hoc Communities of the future there will be no fixed base stations.  Communication between various mobile units will be wireless, but the network structure will be unknown and variable.  Smart Devices (such as smart cars, smart phone, and smart appliances) will communicate with each other without any human interaction. 

4.        There will be a shift in the interaction paradigm used by the database users.  There may be no explicit queries or request/result format to this communication.  Indeed there may be no human involvement at all.  The traditional concepts of transaction and commit seem inadequate to meet the demands of this environment.  In addition, the use of these terms carry certain connotations which are not applicable in this environment.

5.        Due to mobility and variability, conventional approaches to metadata and catalog/directory management are not sufficient.    Ad hoc communities may not even have any explicit metadata available.  This will require on the fly creation of metadata.

 

 

WWW Issues:

1.        Personalization and task-specific information analysis are of paramount importance in helping users grapple with information overload and in matching the limitations of the increasingly popular portable devices. Related issues are profile management (of singular users, groups of users, tasks, and devices), domain customization, and quality of information issues (such as recency, authority, and provenance/pedigree), and presentation, summarization, and classification of the information.

2.        As new data is continuously being introduced, transformed, or becoming inconsistent, methods that allow for dealing effectively with such fluidity of the data need to be investigated. Such methods include, push and pull, subscription and modification, and automated DBAs. Fluidity of another kind relates to dynamic web pages, such as in e-commerce sites.

3.        The “next generation Web” must provide widely available capabilities beyond the keyword search facilities, user interfaces, and information gathering processes that are current available. Thus, research is needed in new paradigms for querying and filtering (e.g., integration of keyword matching with database-like queries, publish/subscribe models, and support for inexact answers or unreliable/inconsistent data), user interaction (e.g., multimodal, visual, natural-language interfaces), and data integration (e.g., webhouses, resource discovery, information-gathering agents, and effective use of metadata).

4.        The transition between a syntactically connected Web and a network of interconnected concepts is underway. Central issues concern helping users create a “mental model” of the network and supporting emerging web communities. Techniques for the management of metadata and ontologies, and for “information ingestion”  (i.e., integrating new knowledge into existing knowledge) are needed.

5.        New challenges in the administration of large-scale information systems include the management of heterogeneous collections, the control of distributed information repositories, and integration with the Web, so as to facilitate widespread and preferential access to the collections (e.g., by conforming to metadata standards, facilitating navigation within the collection, or ensuring data quality).

6.        A wide range of applications can benefit from advances in web database research and at the same time provide invaluable real problems that need to be solved. Examples of such applications include e-commerce,  biology, digital government, collaborative work, and support of processes and workflows.

7.        System issues that arise in connection to the higher level issues previously mentioned include web caching and prefetching, replication, new query evaluation architectures, cost models, data delivery techniques (e.g., push vs. pull, broadcast vs. point-to-point), indexing, multiple-query optimization, and profile evaluation engines.




RECOMMENDATIONS

 

Encourage proposals and funding related to WWW and wireless within IDM.  Specifically include a new focus area, MOBILE and WEB BASED INFORMATION MANAGEMENT based on “new computing paradigms”  such as personalization and interaction shift.  Emphasize research in the following areas:

Wireless Research areas:

1.       Adaptability to Available Resources

2.       Location/time/network/mobility dependent data and queries

3.       Unique requirements of the ad hoc communities and dynamic networks

4.       Performance metrics and benchmarks

5.       Query Processing and Optimization

6.       Mobile/Wireless aware infrastructure including handling of metadata and catalogs

WWW Research areas;

1.       Personalization for querying , presentation, summarization, and classification of information

2.       Large-scale data integration and management of distributed and heterogeneous data

3.       Metadata management

4.       Performance metrics and benchmarks

5.       Query Processing and Optimization

6.       Fluidity management

 

 

 

Support more interdisciplinary and collaborative workshops to foster larger grants.  These should encourage both university and industry participation.

 

 

Discussion Group Participants

 

Co-Chairs:

Isabel F. Cruz

Margaret H. Dunham


Members:

Sigel Adali
Lois Boggess
Philippe Bonnet
Jamie Callan

Hsinchun Chen

Panos K. Chrysanthis

David Cooper

Eric Hanson

Anupam Joshi

Vijay Kumar

B. S. Manjunath

Sharad Mehrotra

Weiji Meng

Bongki Moon

Aris Ouksel

Krithi Ramamritham

Ken Ross

Elke Rundensteiner

Frank Shipman

Prasad Sistla

Nandit Soparkar

Ellen Spertus

Susan D. Urban

Victor Viannu

Nancy Wiegand

Ouri Wolfson

Clement Yu

Stan Zdonik

Shlomo Zilberstein