One Language One Enterprise SEMANTICS DELIVERING ONE LANGUAGE

One Language. One Enterprise. ™ SEMANTICS: DELIVERING ONE LANGUAGE TO THE ENTERPRISE Managed & integrated enterprise information in support of transparency, integration & lower IT costs 2 nd Semantic Technologies for e. Gov 2004 -09 -09 Zvi Schreiber 1 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Goals of the Modern Enterprise • Transparency – – – To management/command chain: Business intelligence To the citizen To other agencies Of events: Business Activity Monitoring (BAM) Agility to change reports and underlying systems • Integration – – Automate processes to reduce cost and increase reliability Integrate with other commands/agencies Adoption of industry standards and web service technology Agility to change & customize • IT efficiency – Consolidation of redundant systems – Productivity of developers 2 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Need for Information Management & Integration • But critical enterprise information resides in hundreds of disparate systems – Each with its own incompatible “data language” for operations, people, items… – Systems change over time – Some systems belong to other commands/agencies • Business goals require managed integrated enterprise information – – – Know where your data is and promote reuse Know what data means in agreed business language Know when data assets change and understand impact Standards for new systems to avoid deepening problem Allow data to be queried across systems as needed • Information management & integration are a non-disruptive flexible way to support enterprise goals • Best practice is “Semantic Information Management”… – Give data objective business meaning so that it may be found, aggregated and used automatically without prior knowledge of its physical format – Inspired by W 3 C Semantic Web vision – Applied today to enterprise data 3 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Semantic Information Management (SIM) Need Technology Catalog data assets, track change, publish Enterprise metadata repository Common business language Ontology model (including rules) for citizen, items, orders, etc. Attach objective business meaning to data Semantic mapping of data schemas to Ontology (map once hub & spoke) Manage information Information management services – discovery, impact analysis, redundancy, classification Integrate information Information integration services – Automatically infer & maintain data transformations and query Virtual enterprise database Federated query – EII (from IBM) 4 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Unicorn’s Architecture for SIM DESIGN Information Resource Managers Unicorn Workbench Integration Knowledge Workers IT Professionals BPM, EAI, SOA Business Intelligence ETL, BI Unicorn Web Interface Enterprise View Interface Unicorn Server Federated Query • Information Management • Information Integration • Information Quality Semantic EII Information Management Portal Semantic Engine™ Unicorn Repository™ COMMON BUSINESS LANGUAGE: Ontology Model Semantic Mappings CATALOG: Metadata l na io II pt O IBM • Mapping • Modeling • Repository Management RUN-TIME MANAGEMENT

Use Cases • Information resource managers – – – Collaboratively capture common language in ontology Catalog data sources and semantically map Apply policy e. g. classification based on common business language Analyze redundancy of data resources Impact analysis • Developers – – Discover data sources Create new schemas based on standard language Infer translation scripts (e. g. XSLT for SOA) Impact analysis • Business analysts – Ad Hoc queries using common business language – Metadata: What does this report mean? – Traceability: Where does this come from? 6 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Case Study – Transparency • The enterprise – Global semiconductor manufacturer – manufacturing planning function • Business drivers – Single view of flow of goods – Resolve data quality discrepancies between different systems causing missed shipments & excess inventory • Semantically different definitions of workweek, product taxonomy • Project nature – A focused two week SIM project mapped different data sources to a single common business language – Sophisticated SQL queries inferred automatically to provide a single business view from physical sources • Results – All discrepancies exposed – Ontology model is reusable and may be extended to a broader view 7 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Case Study – Integration • The enterprise – Leasing US life insurance carrier and IT thought leader – User is large integration hub for all individual business • Business drivers – Quicker integration – More flexibility to update & customize processes to different channels and to change underlying systems – Adopt industry standard messaging (ACORD) • Project nature – Ontology based on ACORD Life standards – All proprietary message formats and all versions and extensions of ACORD messages mapped to Ontology – Transformations planned or generated automatically • Results – – Productivity of development team greatly increased Time to absorb update of ACORD standards down from weeks to hours Changes in underlying systems no longer disrupt processes Mappings shared between this project and other projects 8 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Case Study – IT Efficiency • The enterprise – Leading global investment bank – Purchase sponsored by head of firm-wide IT • Business drivers – – Exposing redundancy & consolidation of 10, 000 databases Developer productivity; reuse Preparation for Sarbanes-Oxley, Basel II Enforcement of naming standards in IT • Project nature – Strategic commitment to SIM for firm-wide roll out in stages – Involvement of multiple specific IT projects • Results – – Increase in developer productivity Increased use of standards naming Redundancy starting to be exposed Traceability of reports and definitions for SOX, Basel II 9 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

From the Analysts • On the common business language approach • On the need for management • On semantic approach to design • On EII runtime • About Unicorn – “Having a common business language can minimize the time and effort (of business process integration), and make these critical business processes much more efficient” (Gartner) – “Having a common business language helps to encourage transparency, quality of information, privacy and security of information” (Gartner) – “The idea of using a common language to… decouple applications from data – is another potential impact on the IT stack. ” (Gartner) – “Large enterprises will refocus on data management as a strategic discipline” (Gartner) – “lack of technologies and products to dynamically mediate discrepancies in business semantics will limit the adoption of advanced Web services” (Gartner) – “EII solutions will be incorporated in portal, EAI, business intelligence, and businessprocess-integration solutions, making EII a key technology in a $7. 5 billion market. ” (Aberdeen) – “Unicorn has a strong enterprise data focus. In trying to 'semantify' enterprise architecture, Unicorn seems to bring some interesting technology and skills to the data and metadata management areas. Among the various players we surveyed, Unicorn seems by far the most enterprise data–centric. ” (Cutter) 10 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Metadata: Catalogue Data Sources • Catalog data sources • Automatically capture technical metadata - e. g. tables, columns and data types • Add descriptors – source, ownership, dependencies etc. • Flexible metamodel based on OMG MOF can accommodate C 4 ISR metamodel • Also catalog and crossreference other IT assets: Applications, servers, etc. 11 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Ontology Model: Common Business Language • Ontology model – Classes (entities – specialization/generalization) – Properties (relationships) – Business Rules (conversions, lookup tables) – Descriptors (definitions, synonyms, …) • Standards – W 3 C OWL – Use/extend ERDs, UML – Use/extend Do. D standard data models 12 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Semantic Mapping • Map each data asset once only as a spoke to Ontology Model hub • Formal semantic mappings capture meaning of data in formal machine-readable form • Flexibility of Mapping – Map all assets: relational, XML, legacy, etc. to same model – Productive mapping in two stages groups (e. g. tables) and fields (e. g. columns) – Attach conditions to mapping 13 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Semantic Discovery • Find all physical data relating to a business concept • Explore indirect relationships – Data which is a subset/superset – Data related by business rules • Apply business policy (e. g. data Privacy regulations) consistently • Reasoning details 14 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Graphical Reporting • Generate reports on any metadata concept • Drill down on graphical reports to determine the source assets for each statistic 15 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Query – Powered by Unicorn + IBM II • Query using common business language • Results using common business language • Seamlessly executed on • Metadata: What does this physical data mean? sources • Traceability: Where did it come • Join across from? multiple sources 16 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Key Benefits to Government • Design: ü Formalize common language (in ontology) – increase communication and avoid ambiguity ü Attach formal meaning to physical data – map once only ü Benefit from W 3 C Semantic Web vision today ü Support Service Oriented Architecture ü Accelerate process automation/integration • Management: ü ü Discovery: Increase reuse of data Identify and eliminate redundancy Increase quality of information (accuracy, clarity) Agility: Accommodate changes in data sources, language and rules • Runtime: ü Query using common language and across data sources ü Traceability of queries to data sources – where did these results come from? What do they mean? ü Dramatic cost savings and flexibility increase over traditional data warehouse technologies ü Use traditional 3 rd party analysis & visualization tools 17 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Challenges • Negotiate common business language – Base on existing standards where possible • Governance – Ensure loading and publication of metadata • Extend traditional modeling & mapping skills 18 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential

Conclusions • Disparate data sources are barrier to transparency and integration & cause wasted IT spend • Best practice is Semantic Information Management – – Catalog data sources (metadata) Establish common business language (ontology) Semantically map physical data schemas to ontology Apply to information management • Classification, redundancy analysis, impact analysis, … – Apply to information integration • Infer translation scripts, turn ontology into a virtual database using federated query (EII) • Start with pain points & expand to entire enterprise 19 © Unicorn Solutions Inc. 04 December 2020 Private & Confidential