IBM InfoSphere MDM v11 Overview - Aomar BARIZ

Editor's Notes

• #5: Finally, I wanted to end our discussion on what is master data and master data management by summarizing a few key trends that may influence your thinking and has influenced IBM’s product direction. 1 – MDM requires information governance to be successful. 2 – Organizations are linking process and data strategies with BPM and MDM. 3 – Implementations are maturing from single domain to multi-domain.
• #7: Masters all domains in a single solution: IBM is unifying best-of-breed technologies that were built to address specific needs of different domains or different technical approaches to MDM. IBM will make this combined technology available to our customers as a single solution to address all their MDM needs across the whole enterprise. Govern Inside & Out: IBM is building in governance capabilities within the MDM technology that are unique to master data, but IBM is also integrating our MDM to work seamlessly with related governance technologies in IBM’s overall InfoSphere governance solution. Designed for Big Data: IBM’s MDM has always been the market leader when dealing with volume and scale. While we continue to push the boundary on volume & scalability, we are also building our MDM to handle the increasing VARIETY of information that is a hallmark of the Big Data trend. Accelerate Time-to-value: As MDM becomes a ‘mainstream’ technology, we continue to look for ways to simplify our software, reducing the time and cost necessary for clients to implement MDM and start getting value from MDM. This comes not just through simplifying the core MDM software itself, but also through providing pre-configured MDM solutions that target specific business problems or are designed for specific applications. (Advanced Catalog Management for Commerce is a great example). Embrace New Era of Computing: New computing platforms are becoming more and more important, both emerging platforms such as ‘converged systems’ (like IBM’s Pure Application systems) as well as already-emerged platforms of mobile computing. IBM MDM is designed to take advantage of these new computing platforms in ways that are valuable for our clients. Rich MDM Ecosystem: Lastly, we realize that the success of our MDM business is not just a function of the software itself, but a function of having a rich and healthy ecosystem of domain experts, business partners, implementation specialists, and thought leaders both inside IBM and outside IBM.
• #8: Masters all domains in a single solution: IBM is unifying best-of-breed technologies that were built to address specific needs of different domains or different technical approaches to MDM. IBM will make this combined technology available to our customers as a single solution to address all their MDM needs across the whole enterprise. Govern Inside & Out: IBM is building in governance capabilities within the MDM technology that are unique to master data, but IBM is also integrating our MDM to work seamlessly with related governance technologies in IBM’s overall InfoSphere governance solution. Designed for Big Data: IBM’s MDM has always been the market leader when dealing with volume and scale. While we continue to push the boundary on volume & scalability, we are also building our MDM to handle the increasing VARIETY of information that is a hallmark of the Big Data trend. Accelerate Time-to-value: As MDM becomes a ‘mainstream’ technology, we continue to look for ways to simplify our software, reducing the time and cost necessary for clients to implement MDM and start getting value from MDM. This comes not just through simplifying the core MDM software itself, but also through providing pre-configured MDM solutions that target specific business problems or are designed for specific applications. (Advanced Catalog Management for Commerce is a great example). Embrace New Era of Computing: New computing platforms are becoming more and more important, both emerging platforms such as ‘converged systems’ (like IBM’s Pure Application systems) as well as already-emerged platforms of mobile computing. IBM MDM is designed to take advantage of these new computing platforms in ways that are valuable for our clients. Rich MDM Ecosystem: Lastly, we realize that the success of our MDM business is not just a function of the software itself, but a function of having a rich and healthy ecosystem of domain experts, business partners, implementation specialists, and thought leaders both inside IBM and outside IBM.
• #10: PURPOSE: Identify how this offering will be validated with customer and business partner needs SOURCE: Marketing/Release Lead/Green Thread Lead/ Product Owner (Agile)/Business Development (for Business Partners) CHANGE HISTORY: Originally introduced in June 2008 as a “pilot”. Pilot designation removed in Version 5.4 (July 2009) Dec 09: Added specific reference to Business Partners Provide a list of the customers and/or business partners the Offering Team/Green Thread Team is working with, or plans to work with to ensure that this offering meets their needs. Also describe how the offering team plans to solicit input from the customer/business partner, the customer checkpoints, and the plan for incorporating customer feedback into the offering release processes. The project team documents its plans for Early Programs via the SWG Quality Plan template, Customer Early Program Feedback Summary Tab.
• #11: Probabilistic Search – re-use the sophisticated matching capabilities for searching as well Application Toolkit - create sophisticated master data UIs or embed MDM capabilities in end user UIs, including BPM RDM - Provides the governance, process, security and audit control for managing reference data as an enterprise standard, resulting in fewer errors, reduced business risk and cost savings Building on the unified packaging introduced with IBM InfoSphere Master Data Management V10.0, V10.1 contains significant enhancements that include: Master Data Governance Master Data Policy Administration – Define and create master data policies to ensure trusted data quality for consumers of master data Master Data Policy Monitoring – Provides reports and dashboards to track the uniqueness and completeness of the golden record and measures the consistency between the data sources and the golden record Master Data Policy Enforcement – Ensures noncompliant data is remediated, critical data changes are reviewed and approved, and master data is appropriately matched, linked, and collapsed Business Process Management (BPM) integration capabilities and sample workflows – Straightforward mechanisms to create processes (workflows) that govern data steward-oriented and other tasks. Master Data Management Enhanced probabilistic matching and searching – Probabilistic capabilities are expanded to include probabilistic search for the InfoSphere MDM Server technology and use of the latest probabilistic matching engine. Advanced Rules Management – Enables business users to leverage a single interface for product creation, rules authoring, and rules association through an integrated solution with WebSphere ODM. Helps drive product personalization, advanced bundling and configuration, complex offers, and more. Advanced Catalog Management - An accelerated solution with an out of the box data model, business process workflows, and integration components tailored for WebSphere Commerce. Includes advanced capabilities for managing eCommerce catalogs supported by WebSphere Commerce Additional support for delimited file formats for Adaptive Services Interface – After introducing of the Adaptive Services Interface in v10.0 its support for XML message formats has been augmented with support for delimited file and messaging formats, thereby further expanding the number of integration points that can be supported and removing the need to “code” the required interfaces. Enhancements to the InfoSphere Master Data Management Application Toolkit – With additional widgets, the Application Toolkit can be used to create sophisticated master data UIs or to embed MDM capabilities in end user UIs, e.g. in CRM or Call Center applications.
• #14: Service patterns are defined in the core bundles. These patterns include definitions for business objects, lookup tables, business proxies, etc. These same patterns are then employed by customers who extend the core features of MDM with their own customizations Central framework is agnostic regarding the where a service provisioned. There is a pattern of look-ups that is employed to get the correct bundle for any given service Use of OSGi to allow for a deployment of InfoSphere MDM that is more modular/granular Separation of core InfoSphere MDM code (the IBM code) and the extensions (the customer’s code) as components (or “bundles”) Ability for customers to create multiple bundles such that separate development teams can work more independently Patching can occur more localized; IBM patches against the core code bundle; the customer’s patches against the extension bundle(s) 2 scenarios: Customer would have to take ibm EAR out of box into workbench, then take own code and redeploy everything – net new EAR for every upgrade and all patches Every customer said this is painful Organize development teams better and segregate extensions/customizations – can have different bundles, separate code over bundles Examples – cardinal health – have multiple teams, by domain, customer data use cases; product data use cases – can now more easily work independently where appropriate, when working cross domain still work together – (data model changes go cross – really only one data model) – Use of OSGi to allow for a deployment of InfoSphere MDM that is more modular/granular Separation of core InfoSphere MDM code (the IBM code) and the extensions (the customer’s code) as components (or “bundles”) Ability for customers to create multiple bundles such that separate development teams can work more independently Patching can occur more localized; IBM patches against the core code bundle; the customer’s patches against the extension bundle(s) Customer would have to take ibm EAR out of box into workbench, then take own code and redeploy everything – net new EAR for every upgrade and all patches Every customer said this is painful Organize development teams better and segregate extensions/customizations – can have different bundles, separate code over bundles Examples – cardinal health – have multiple teams, by domain, customer data use cases; product data use cases – can now more easily work independently where appropriate, when working cross domain still work together – (data model changes go cross – really only one data model) –
• #16: Examples of Hybrid scenarios Prospect/customer use case Source/consuming systems Mainframe can only contribute info, can’t consume updates – technology is too expensive to create that integration back – not worth it CRM – can consume M&A - I just bought a company that has X systems, bring their info into virtual side for comparison, but can’t persist yet because you may not own data yet or post acquisition but systems aren’t ready to consume Cigna – customer service – moving from account/policy centric to customer perspective - feed customer service portal Dell – started with identifying individuals cause they didn’t know who they were, duplication, important when I interact to know where they are: identification of individual = first step in journey; over time, want much richer customer profiles, adding info such as privacy preferences (can we call you, email), demographic view to broader view – account management view – know if you are assigned to that account - – getting to full set of attributes is not easy – needs people to agree, get through IT structure Allstate – Wells Fargo – Matching tech is why IBM bought initiate - New customer – start quick (identify specific use case, get people to agree, approach different LOBs to say you still own data but I’ll take your data and give other areas the best view of the data), as LOBs become comfortable with data overtime, I can now grow over time – low risk – journey – flexibility to do either/both when appropriate – can make journey if needed – - Go with IBM because you aren’t locked into certain architecture – make it easy for you to change your IT based on your business requirements vs. other way around Hybrid prospect/customer – thin/think; some systems can consume, some can only provide – Mainframe can only contribute info, can’t consume updates – technology is too expensive to create that integration back – not worth it CRM – can consume M&A - I just bought a company that has X systems, bring their info into virtual side for comparison, but can’t persist yet because you may not own data yet or post acquisition but systems aren’t ready to consume How we are different from INFA – we have done it more; experience, reference V11 Harmonized Terminology Engine Co-Residence MDM Server and Initiate MDS engines within same WAS deployment Simplified deployment and management (based on OSGi) Unified MDS/MDMS Workbench for configuration and customization Unified MDS/MDMS installer Integration with IBM Support Assistant Data Collector for both MDM SE/AE Modularity - Use of OSGi to allow for a deployment of InfoSphere MDM that is more modular/granular Separation of core InfoSphere MDM code (the IBM code) and the extensions (the customer’s code) as components (or “bundles”) Ability for customers to create multiple bundles such that separate development teams can work more independently Patching can occur more localized; IBM patches against the core code bundle; the customer’s patches against the extension bundle(s) Enforce versioning of third party code components
• #17: Examples of Hybrid scenarios Prospect/customer use case Source/consuming systems Mainframe can only contribute info, can’t consume updates – technology is too expensive to create that integration back – not worth it CRM – can consume M&A - I just bought a company that has X systems, bring their info into virtual side for comparison, but can’t persist yet because you may not own data yet or post acquisition but systems aren’t ready to consume Cigna – customer service – moving from account/policy centric to customer perspective - feed customer service portal Dell – started with identifying individuals cause they didn’t know who they were, duplication, important when I interact to know where they are: identification of individual = first step in journey; over time, want much richer customer profiles, adding info such as privacy preferences (can we call you, email), demographic view to broader view – account management view – know if you are assigned to that account - – getting to full set of attributes is not easy – needs people to agree, get through IT structure Allstate – Wells Fargo – Matching tech is why IBM bought initiate - New customer – start quick (identify specific use case, get people to agree, approach different LOBs to say you still own data but I’ll take your data and give other areas the best view of the data), as LOBs become comfortable with data overtime, I can now grow over time – low risk – journey – flexibility to do either/both when appropriate – can make journey if needed – - Go with IBM because you aren’t locked into certain architecture – make it easy for you to change your IT based on your business requirements vs. other way around Hybrid prospect/customer – thin/think; some systems can consume, some can only provide – Mainframe can only contribute info, can’t consume updates – technology is too expensive to create that integration back – not worth it CRM – can consume M&A - I just bought a company that has X systems, bring their info into virtual side for comparison, but can’t persist yet because you may not own data yet or post acquisition but systems aren’t ready to consume How we are different from INFA – we have done it more; experience, reference V11 Harmonized Terminology Engine Co-Residence MDM Server and Initiate MDS engines within same WAS deployment Simplified deployment and management (based on OSGi) Unified MDS/MDMS Workbench for configuration and customization Unified MDS/MDMS installer Integration with IBM Support Assistant Data Collector for both MDM SE/AE Modularity - Use of OSGi to allow for a deployment of InfoSphere MDM that is more modular/granular Separation of core InfoSphere MDM code (the IBM code) and the extensions (the customer’s code) as components (or “bundles”) Ability for customers to create multiple bundles such that separate development teams can work more independently Patching can occur more localized; IBM patches against the core code bundle; the customer’s patches against the extension bundle(s) Enforce versioning of third party code components
• #22: Level 1 – social media profile in MDM (today ’ s capability) – SM contact points, preference profile, …. What else? Level 2 – integration with Big Data to aggregate SM into MDM facts – intent to purchase, sentiment towards X, ….. V11 – the landing point for new insights is MDM, so they can operationalize the insights from Big Data
• #23: Our product management, engineering, marketing, CTPs, etc, etc teams have all been working together to help to better understand the big data market. We ’ve done surveys, met with analysts and studied their findings, we’ve met in person with customers and prospects (over 300 meetings) and are confident that we found market “sweet spots” for big data. These 5 use cases are our sweet spots. These will resonate with the majority of prospects that you meet with. In the coming slides we’ll cover each of these in detail, we’ll walk through the need, the value and a customer example.
• #24: Turn MDM Repository into an out of the box Text Analysis tool Automatically discover linkages between text and relevant MDM entities. Enhance detection of relationships between entities Enhance entity resolution from the evidence hidden within the documents. Enrich knowledge base by adding additional information to MDM records Data model agnostic – can be for any domain – ENTERPRISE EDITION ONLY Traditional internal data sources – integrate into consolidated view to downstream systems Potential information about customers lost out in unstructured, xls, reports, email, social blogs – might be in ECM, used to be unavailable to MDM With EUTC, able to take info about customers and use it to exploit Go to one location to see information from traditional and non traditional sources related to this person See connections – Text analysis tool – what emails, what information in unstructured is linked to customers, degree of certainty that it is linked to this entity, relationship link to sources Now I know more about this customer, conversations - tech support – know customer, name, (demographics), know email conversations – know experience, so the next tech support person is more prepared for interaction If negative, can say, I understand you had problems last time you reached out, is that what you are calling today or did it get resolved If positive, can say, did this fix work Public safety – structured data across multiple sources across the world, no connection around address so not able to say there is a link, If have 2 names, 2 addresses in email, typical text analytics will not find links between attributes, combine MDM as part of entity resolution – create an entity from attributes Analysts – given links made by EUTC, enter lead info, I want to know everything about jennifer reed from all my sources, results come up, query, click on entities, see relationships/connection between structured and unstructured – INVESTIGATIVE TOOL Link but don’t automatically resolve – don’t want to update master record from unstructured sources, but want to have link. Unstructured is not as trusted so shouldn’t update master record. User could update master record based on info found in unstructured doc PMe offers attribute level scoring for us to be able to get granularity on how close information in document relates to information in hub - if rare, more sure – if commonly occurs – then not likely a link -- need granularity of proprability to know likely that this is a link – because unstructured is less trusted Identity Insight only does similarity score but can’t say it is a common address or common name (high occurrence/appears a lot in repository) – doesn’t know probability that it occurs –
• #25: Public safety – structured data across multiple sources across the world, no connection around address so not able to say there is a link, If have 2 names, 2 addresses in email, typical text analytics will not find links between attributes, combine MDM as part of entity resolution – create an entity from attributes Analysts – given links made by EUTC, enter lead info, I want to know everything about jennifer reed from all my sources, results come up, query, click on entities, see relationships/connection between structured and unstructured – INVESTIGATIVE TOOL Link but don’t automatically resolve – don’t want to update master record from unstructured sources, but want to have link. Unstructured is not as trusted so shouldn’t update master record. User could update master record based on info found in unstructured doc
• #26: Augment traditional product information with dynamically derived product traits based on web and social media feedback Can use this additional information to update marketing, offers, address technical issues, change product specs, change packaging etc. This is an example where the master record could be updated/enhanced with insight from unstrcutured text.
• #28: IBM combines proven MDM technologies with new innovations from IBM Research to enable healthcare organizations to find and perform analytics on groups of patients who are similar in terms of their clinical information. Key Functions HL7 integration with Healthcare specific Natural Language Processing Text analytics based on healthcare vocabularies, including SNOMED, LOINC, RxNorm Health record de-identification Patient similarity service against a longitudinal patient record Clinical trials cohort selection application UI Key Benefits Ease of use for researchers Quick time to value Better and broader results from the unique advanced analytics approach Built on gold-standard MDM/EMPI technology
• #34: IBM IOD 2012 07/15/13 Drury Design Dynamics
• #35: Translation – different language UIs to support country implementations (big in europe) Easier install – had to install 2 things prior, now just install 1 thing Security enhancements, attribute level security – greater control over data ownership Global search – search for reference data across sets – want to find a ISO code, search for that name using global search; or if looking for country code = US – find all sets where code is present Slide 18 -= screen shot Manage changes Advanced hierarchies Integration with IIS