Intelligent Information Systems 5. Mediated Sytems Gio Wiederhold EPFL, April-June 2000, at 14:15 - 15:15, room INJ 218 Schedule Presentations in English -- but I'll try to manage discussions in French and/or German. Integrated approach, drawing from concepts in databases, artificial intelligence, software engineering, and business principles. 1. 13/4 Historical background, enabling technology:ARPA, Internet, DB, OO, AI., IR 2. 27/4 Search engines and methods (recall, precision, overload, semantic problems). 3. 4/5 Digital libraries, information resources. Value of services, copyright. 4. 11/5 E-commerce. Client-servers. Portals. Payment mechanisms, dynamic pricing. 5. 19/5 Mediated systems. Functions, interfaces, and standards. Intelligence in processing. Role of humans and automation, maintenance. 6. 26/5 Software composition. Distribution of functions. Parallelism. [ww D.Beringer] 7. 31/5 Application to Bioinformatics. 8. 15/6 Educational challenges. Expected changes in teaching and learning. 9. 22/6 Privacy protection and security. Security mediation. 10.29/6 Summary and projection for the future. • Feedback and comments are appreciated. 7/26/2016 EPFL5 mediators 2 Architectures & Communication History Presentation Information Aggregation Access, Select Data Source Printed reports Application Computatio I-O code Local Storage Function: ‘mainframe’ 7/26/2016 terminl Application Computation SQL for A&S Data Base smart terminal Work station Information Minicomptr Application Computation Select FTP File Storage CORBA file server client server EPFL5 mediators Object Struct. Server Storage User Workst. Information Aggregation SQL, ... for A&S Distr. Sources mediated 3 The world wide information network and its participants _ …. _ External: …. sources and / or sinks _ …. _ _ …. …. _ data, meta-data, knowledge _ …. _ _ …. _ _ …. _ …. …. …. _ …. _ …. _ …. _ _ …. _ …. _ …. _ …. _ …. _ …. …. …. 7/26/2016 EPFL5 mediators _ …. _ …. Internal: transformers and memory. 4 Mediation moves services into the network applications A4 A3 A2 A1 A5 A6 integrators a. I2 I1 mediators network b. M1 c. d. M2 e. wrappers D1 W2 W1 D2 D3 D4 W3 D5 D6 datasources History 7/26/2016 EPFL5 mediators 5 Understand the Architecture for Information Technology: Component Classification Customers Customers Customers Customers Customers Services Services Services Sources Sources Sources 7/26/2016 EPFL5 mediators 6 Transform Data to Information Application Layer Mediation Layer decision-makers at workstations value-added services Foundation Layer data and simulation resources 7/26/2016 EPFL5 mediators 7 Service layer Between the minds that plan and the hands that work, there must be a mediator. [Fritz Lang, intertitle in 'Metropolis' UFA 1926] Decision making (DM) Multiple domains ! Customer Service MEDIATION Resource access 7/26/2016 Shared software, standards ? EPFL5 mediators 8 Example in Health Care . Health Care Planner Will the Clinic loose Money? Investment domain Patient Care domain Age Profile Service Operations Patient Volume Growth 7/26/2016 Loan Interest Bond Sales State Support Specifications for the components Customer models Customers Customers Customers Customers Customers Services Services Services Sources Sources Sources 7/26/2016 Catalogs Content & Methods Metadata EPFL5 mediators 10 Value-added intermediate services 1 Needs Technologies extant and new Describe customer Build interpretable workflow model with model meta-specifications for selection Discover new resources Monitor and index public metadata, describe resource capabilities, contents & methods Select relevant resources Match available metadata and indices of resource contents to leaf nodes in the customer model Easy access to resources Wrap resources to make them compatible, exploit wrapper templates, skip unavailable sources Filter out excessive data Filters attached to the customer model; balance relevant volume and precision 7/26/2016 EPFL5 mediators 11 Modeling: sources • Models provide abstractions • abstractions represent a point of view • Models of databases are schemas and E-R models • well established • constraints - references, uniqueness • scopes remain implicit • Information systems have meta-data • XML has DTD’s • under discussion, still limited Focus on resources Meta data 7/26/2016 EPFL5 mediators 12 Value-added intermediate services 2 Needs Identify articulations * Match level of detail * Integrate information Omit redundant data, documents Reduce customer overload Inform customer 7/26/2016 Technologies, extant and new Matching of related concepts, use articulation rules to match nodes Automatic abstraction to match sources at articulation points within the customer model Attach data instances to articulation points, combine elements , link to customer model Match data for content, omit overlap, report inconsistencies in overlapping sources Summarize according to customer model, rank information at each level Present information according to model hierarchy, consider bandwidth EPFL5 mediators 13 Integrate via Mediation • User application – Workstations • Mediator Services – Expert-owned nodes – Composable • Data sources – Remote primary and byproduct services – Simulations 7/26/2016 EPFL5 mediators 14 Mediation for Quality S= source reliability C= confidence T= Estimates: C1= 5+_1 T1=100+_160 S1 7/26/2016 Customer Model f(S,C,T) BEST= low cost rapid response Assessments: reliable delivery S1=.8 S2=.9 S3=8 trustworthiness C2= 8+_1 T2=70+_30 C3= 10+_1 T3=50+_80 S2 S3 EPFL5 mediators 15 Functional Layer User interface Human-computer Interaction Applicationspecific code Service interface MEDIATION Resource access interface Wrappers Domainspecific code Sourcespecific code Real-world interface 7/26/2016 EPFL5 mediators 16 Productivity • Data is NOT information – Information supports decision-making – Best among a limited number of choices • Selection of best sources – Locate all, use the best, most relevant • Integration of diverse sources – Time synchrony, spatial synchrony, topic matching – Abstraction, ranking, pruning • Intermediate processing to overcome – Disintermediation between source (author) and destination (customer) 7/26/2016 EPFL5 mediators 17 Functions in Mediation of Data • Delivery to client • Summarization or determine exceptions from expected values or trends • Omission of replicated or known information • Integration of data from diverse domains • Resolution of scope mismatch transistors semiconductors • Abstraction to match levels of granularity • Conversion to compatible protocols and representations • Assessment of quality of diverse sources • Acess to sources via wrappers 7/26/2016 EPFL5 mediators 18 Mediator Design Principle Transform Data into Information Match Customer Model Hierarchical to Resource Model General network and maintain models (large fraction of cost) 7/26/2016 EPFL5 mediators 19 Assigning maintenance responsibility a. Source data quality – supplier database, files, or web pages b. Interface to the source – Sources wrapper, supplier or vendor for supplier c. Source selection – expert specialist in mediator d. Source quality assessment – customer input to mediator Services e. Semantic interoperation – specialist group providing input to the mediator f. Consistency and metadata information – mediator service operation or warehouse g. Informal, pragmatic integration – client services with customer input h. User presentation formats – Customers client services with customer input 7/26/2016 EPFL5 mediators 20 Allocation Flexibility User Interfaces Application C Provider of Mediator M 7/26/2016 Application I Provider of mediator N M Copy- if high intensity of interaction with 1. Application (M2) 2. Resources (N1,2) 3. Processing (M1) Mediators are only code Application B M2 HPC N M1 DB P N 2 Databases N 1 DB Q EPFL5 mediators DBS R 21 Central Solutions do not Scale What works with 7 modules and one person in charge fails when we have 100 and need a committee Changes in resources affect the intermediary modules 7/26/2016 EPFL5 mediators 22 Growth through Reuse Gio Wiederhold. 1995 New Application Prior & Revised Mediators Extended Data Resources 7/26/2016 EPFL5 mediators 23 Linear O(n) Cost of Growth else O(n2) • Data changes only affect some mediators; only in their domain • Mediators can 1. supply old information to n-1 prior applications 2. provide better information to the new application 3. be partially or completely reused • New applications, using the new data, can be developed and inserted dynamically 7/26/2016 EPFL5 mediators 7 2 24 Architecture instances Applications . . . . Mediators . . . . . . Resources . . . _ …. …. . _ …. …. . _ …. …. . include computational resources 7/26/2016 EPFL5 mediators 25 Environmental Restoration at INEL Undoing 50 years …. Overall Architecture MSL [Stanford] OQL [ODMG] MQL [ISX] QEM OEM QEM other mediators wrapper OEM OEM QEM QEM OEM mediator QEM OEM OEM QEM CORBA OEM QEM QEM wrapper wrapper wrapper Many projects many sources ERIS LOCKHEED MARTIN 7/26/2016 7/26/2016 IEDMS ISX - Stanford Univ. Idaho National Engineering Laboratory EPFL5 mediators 26 Mediation to Implement Feedback in Training David Maluf, Priya Panchapagesan, Ted Linden Another task of mediators, prior to integration MIFT Abstraction Abstraction to match levels of granularity 7/26/2016 EPFL5 mediators 27 Mediation Feedback: Playback or Graph Commanders User Interface Trainees Observers Training Analysts Developers UI in Java Application Layer Standards in KQML Objectives Mediation Layers Tasks Stanford Mediators with rules in CLIPS I.D.A Wrapped Simulation Resources 7/26/2016 Wrappers in C/C++ Janus EPFL5 mediators SimNet 28 MIFT . Result . Analyses: • Force ratio • Losses • Area gain Exercise Simulator Type 7/26/2016 EPFL5 mediators 29 A mediator Is not just static software Application Interface Changes of user needs Software & Peo ple Owner/ Creator Maintainer Lessor - Seller Advertisor Models, programs, rules, caches, . . . Resource Interfaces 7/26/2016 EPFL5 mediators Domain changes Resource changes 30 Alternative solutions • A Super Database – unwieldly – obsolete before it is established • Distributed, free standing databases (today) – awkward for sharing information (much knowledge derives from the intersections) – hyperlinks and shared references allow navigation • Distributed databases & a single standard allowing interoperation – standards follow progress, cannot lead it • Distributed databases with published formats – requires rapid adaptation to keep up with resources (but the number of resources per project will be limited) with mediators to isolate projects from resources 7/26/2016 EPFL5 mediators 31 Mediator generation • Today mainly handcraftrd – tools at INEEL, using CLIPS for rules – many partial tools in industry (TUXEDO, M.. ) • TSIMMIS – from wrapper descriptions and OEM formats • Blades using RDF [Melnik: DigLib meeting 2000] 7/26/2016 EPFL5 mediators 32 Coverage of Current I3 Efforts (web,schema searching) ) for relevance to customer Maintenance (rule technology?) Integration over sources Wrapping (syntactical heterogeneity) :-( Databases / Web / Text / Simulation :-[ :-( :-) for cooperation for multiple domains :-( :-| Security Mediators :-( :-[ History (auto linking) :-| Caching / Facilitation :-( Abstraction :-[ :-) Discovery | ( ] Good progress / active research / related work / poor coverage :-[ :-) :-) Some Research Issues • Performance – An intermediate layer induces cost (vs. benefits) caching, up to warehousing, – optimizing the balance based on rate-of-change • Semantic ambiguities – Words mean differing things in differing contexts – Integration over all contexts reduces precision Ontology algebra (topic in presentation # 8) • Functions for decision support – temporal algebra, planning, assessing alternatives Simulation as a service (topic in presentation # 10) 7/26/2016 EPFL5 mediators 34 DM support is disjoint Sciences Technologies do not interoperate Planning Science Distribution 7/26/2016 EPFL5 mediators extensions to move to networked support are also disjoint 35 Summary To sustain the trend 1. The value of the results has to keep increasing precision, relevance not volume 2. Value is provided by experts, encoded as models of diverse resources, customers Problems to be addressed mismatches quality Clear models temporal extensions maintenance } 7/26/2016 EPFL5 mediators 36