Towards an Integrated Data Management Platform for the Web

Towards an Integrated Data Management Platform for the Web Stefan Deßloch, Eileen Lin, Nelson Mattos, Portals are the latest technology de- Dan Wolfson, Kathryn Zeidenstein veloped to access, and present personal- ized information from multiple sources to Towards an Integrated Data end-users. The critical need to find and access information at a level below that Management Platform for the Web covered by processes and transactions has led to different technologies, such as web crawling and text categorization in sup- port of portal technology. 1 Introduction tegration problem and that allows us to All of these technologies are primari- define a set of key requirements for defin- ly focused on delivering information to What is role of database in a world con- ing an information integration platform people, rather than other applications. nected by the Web? Database technology for the Web. The style of interaction is predominantly has its roots in, and can measure its suc- Third, we present a survey of current a user-initiated request/response in which cess by, traditional OLTP and decision research and development efforts from the request is often expressed through a support applications supporting business. IBM, which are, in total, working toward form or URL, and the response is provid- In light of the huge success of RDBMS in the goal of solving the complex problems ed as rendered HTML delivered to a web such traditional business applications, is and requirements that are described in the browser for the user to interpret. there a chance for similar success for set of requirements. The success and ubiquity of these In- RDBMS in a world that has significantly We conclude by reiterating that there ternet technologies has opened the door shifted its focus toward reliance on the is an extensive amount of work to be done for new forms of commercial exploita- Web for connecting people, processes, over a diverse set of technical areas to tion. Perhaps this is most evident within and data? provide a robust solution. the emerging area of Web services. Web As a consequence of the Web's roots in documents and rich media delivery, it services simplify application-to-applica- has a completely different data and pro- 2 Background tion interactions by defining and express- gramming model than that employed by ing these interactions using XML. The From the first tentative steps towards in- traditional RDBMSs. The RDBMS has Simple Object Access Protocol (SOAP) formation sharing over the internet in responded to the diversity of Web content was originally proposed by IBM and Mi- 1989, the World Wide Web has become a by adding object-relational capabilities. crosoft and is now being processed as the mainstay of communications and com- However, our view is that this is only a XML Protocol by the World Wide Web merce. Communications of course started first step in the evolution of the RDBMS Consortium (W3C) [W3CSOAP1 2001, with intertwined static HTML documents for the Web. Just as the Web has broken W3CSOAP2 2001]. served up by HTTP servers to form webs down barriers between businesses and Typically, SOAP messages are sent of topical information to simple client other organizations, we believe that core over HTTP between a service requestor applications called browsers. Over the Internet standards and specifications pro- and service provider. This simple proto- years, the static HTML has been com- vide the RDBMS with the tools to enable col facilitates business-to-business inter- bined with more dynamic content to form it to solve the largest problems facing IT actions; for example a product distributor today: Integrating information within and a completely new kind of interactive me- dia in which data from multiple, hetero- may offer a Web service interface allow- between enterprises. Indeed, this enter- ing customers to place or inquire about prise information integration layer must geneous data sources may be combined, displayed, and, in a limited way, queried. orders. Although Web services may be work closely with, and provide integra- As the communications facility of the used for performing such business trans- tion points to, software higher up the Web has rapidly evolved, so has the real- actions, they also serve as a popular technology stack, such as the application ization of commercial potential. The first mechanism for requesting dynamic infor- server and business process integration. and primary commercial use of the Web mation. For instance, companies today To understand why we believe in this was for dissemination of marketing and offer Web services that provide stock expanded role for the RDBMS – that of support information. It is almost unthink- quote information, current weather in dif- an information integration platform – and how we got to this point, we present the able for a company today to not have ferent cities, status of airline flights, or following main points: some kind of web presence to proclaim its the status of express mail packages. First, a brief historical perspective existence to the public. These kinds of information services pro- about the evolution of business on the Through CGI, ASP, and server-side vide a wide-ranging source of heteroge- Web from its roots as a way to present in- Java, information delivery was augment- neous content, content that may reflect a formation to people, to its latest incarna- ed by the ability for customers to actually real-time state as well as more static in- tion as a way to enable program-to-pro- transact commerce on the Web – in ever formation. With such a rich variety of gram communication for information ex- more sophisticated ways, businesses have data sources, we must architect a way to change and for providing services. found ways to employ the web to both harness these new information resources Second, we present a scenario that il- make money and to reduce the cost of do- and integrate them with, perhaps, more lustrates the scope of the information in- ing business. traditional sources of information. Datenbank-Spektrum 2/2002 5 Towards an Integrated Data Management Platform for the Web 3 Changes Ahead for Data as the information is received.