Infrastructures and Architectures post the .com hype and crash Alexander Voss National Centre for e-Social Science alex.voss@ncess.ac.uk www.ncess.ac.uk 25th Sept., 2006 The .com Bubble Ca. 1997-2001 Emergence of the Web Applications mainly around – Simple, one-way content delivery models – Company presentation and advertising – Customer relationship management New, unconventional business models emerging Sector awash with money for: – Internet and network infrastructure (e.g. Worldcom) – Internet tools (e.g. Netscape) – Consumer websites and portals (e.g. Yahoo) Only few business models turned out to be successful, leading to the ‘.com bubble burst’ ca. 2001 A healthy ‘shake-out’ for the industry? 25th Sept., 2006 What’s different today? Growth of the Web/Internet has continued… Technologies have matured, interoperability increased But, crucially, the Web has started to become a platform for more sophisticated applications Much wider range of uses, partially replacing desktop functionality and emphasising collaborative aspects of computing Enabled by: – Advent of XML as extensible data interchange format in server-toserver communication – Availability of components, application server containers and established programming models – Higher level of abstraction as a starting point for systems development 25th Sept., 2006 Web 2.0 and Mashups: .com coming of age? Web 2.0: cover term for technologies enabling rich user experiences and complex functionality Mashup: trading and assembly of functionality through public APIs – http://www.programmableweb.com/ Web Application Hosting – – – – SalesForce.com: online CRM system Google writely: online text processor Del.ico.us: online, collaborative bookmarking YouOS, Microsoft’s Live.com, Yahoo 360° : desktop in a web browser – MySpace.com 25th Sept., 2006 Web 2.0 and Mashup Examples http://www.connotea.org/ http://www.scanbuy.com/ http://tvmap.thomasscott.net/ http://lovejoy.nerc-essc.ac.uk:8080/Godiva2 25th Sept., 2006 Components and Services The programming models behind Web 2.0 and Mashups are the same used for building enterprise applications, i.e. The boundaries between ‘the web’ and everything else are becoming difficult to define. Components & services also used within, even what is sold as an “application” may consist of a configuration of individually developed components Inversion of control, programming to interfaces and aspect orientation make new degrees of independence between modules possible and allow a true ‘pick-n-mix’ approach to application assembly, sourcing from a ‘best of breed’ market. 25th Sept., 2006 Some emerging themes User-customer-provider relationships are becoming increasingly complex and dynamic, functionality arises from combinations of independent parts – – – – – – Security and privacy in dynamic assemblages Provenance of data, auditing its processing Demarcation of areas of responsibility and locating problems Maintaining technological and operational alignment Managing usage, avoiding abuse Increased importance of and changing modes of standardisation? 25th Sept., 2006 Relevance: yet another bubble? Amazon claims to have made 28% of its turnover through its programmable interfaces in 2Q05: $490M (source: iX Magazin) eBay processed 8 billion web service requests in 4Q05 and counted 1900 applications using its services (source: iX Magazin) 25th Sept., 2006 Wider Relevance of Grid/eScience Many future applications will require functionality being developed in the Grid/eScience community, may problems are the same: – security/privacy requirements (MySpace - CancerGrid) – provenance and conditions of use (Flickr - UK data archive) eScience and other applications are starting to inform each other, Grid technologies and the Web Service world are beginning to merge 25th Sept., 2006