Transcription

MDM Alert The MDM InstituteIndependent. Authoritative. Relevant.Field Report: Software AG's webMethods OneData for RDMMonday, November 26, 2012To automatically subscribe/unsubscribe, enter your e-mail address on the form at this link.You may also unsubscribe by sending a blank e-mail to mailto:[email protected] with the subject: UNSUBSCRIBE.Why is “Reference Data Management” So Important?For many software vendors and Global 5000 enterprises, Reference Data Management (RDM) is arelatively new offspring of Master Data Management (MDM) functionality. RDM provides theprocesses and technologies for recognizing, harmonizing and sharing coded, relatively static datasets for “reference” by multiple constituencies (people, systems, and other master data domains).Many current MDM vendors such as IBM, Informatica, Orchestra Networks, and SAP have repurposed their MDM hub functionality to manage reference data as a special type of master data.For Software AG, RDM is not something new as most of their customers, such as Avon, Nokia, andWellpoint have been mastering reference data in OneData for years.Such a system provides governance, process, security, and audit control around the mastering ofreference data. In addition, RDM systems also manage complex mappings between differentreference data representations and different data domains across the enterprise. Mostcontemporary RDM systems also provide a service-oriented architecture (SOA) service layer forthe sharing of such reference data.Prior to the availability of commercial RDM solutions, organizations built custom solutions usingexisting software such as RDBMS, spreadsheets, workflow software (business processmanagement or BPM) and other tools. Such systems often lacked change management, auditcontrols, and granular security/permissions. As a result, these legacy solutions have increasinglybecome compliance risks. Because reference data is used to drive key business processes andapplication logic, errors in reference data can have a major negative and multiplicative businessimpact. Mismatches in reference data impact on data quality affect the integrity of BI reports andalso are a common source of application integration failure. Just as businesses no longer buildtheir own CRM, ERP, and MDM systems, so too are organizations beginning to acquirecommercial RDM solutions, which can be easily tailored or configured and have the full ongoingsupport of a major software vendor.Within the realm of commercial RDM solutions, there are two main families: "multi-domain RDM,"and "real-time RDM". "Multi-domain RDM" solutions are non-industry specific solutions that canspan functional areas (finance, risk and compliance, human resources) and content types (ISOcountry codes, and other non-volatile reference data to be mastered and shared). "Real-timeRDM" is typically a very high performance solution for use in the capital markets industry (brokers,asset managers, and securities services firms) as well as command and control military/intelligencemarkets.Increasingly, many large enterprises have begun to make RDM their initial test case or proof-ofconcept for their MDM evaluations. As a consequence, MDM vendors are rushing to market RDMsolutions to apply an MDM approach for centralized governance, stewardship and control.Cognizant, iGATE Patni, Kingland Systems, Wipro Technologies, and other systems integratorswill move into the "securities master" market (some by OEMing of Informatica and IBM MDM). By2015, pervasive, low cost RDM will be commoditized via the efforts of Microsoft and Oracle as

these vendors provide low or no-cost RDM solutions as part of their software families. Moreover,as many large enterprises have begun to make RDM their initial test case or proof-of-concept fortheir MDM evaluations, the vendor community is responding by providing easier-to-manage entrypoints into RDM use cases using either existing MDM platforms or purpose-built RDM solutionswhich use MDM as their foundation. Clearly, managing “simple” reference data will prove to be akey sales entry point for large enterprises and their MDM vendors. Additionally, RDM can beexpected to become a "ramp up" point of entry for many organizations planning for CUSTOMER,PRODUCT master and other domains, as well as an entry point into master data governance.Moreover, RDM is needed in both operational and analytical MDM use cases where the capabilityis often used to provide attributes, hierarchies and KPIs.During 2013-14, we believe a great amount of current and next-generation commerce will befacilitated by on-premises and cloud-based RDM solutions to support both "private" and "public"reference data. "Public" reference data is what many people typically think of when they considerreference data. Public reference data is based on standards where overall consistency is a primarygoal. Examples of public reference data include industry standards (GS1 GPC), national standards(FIP 10-4, US Census MSA/CSA), International Standards (ISO, ISIC), and data from vendors(Bloomberg, D&B, S&P). “Private” reference data is used to maintain consistency when doingbusiness with external parties. Examples of private reference data include financial andorganizational hierarchies, and employee organizational structures. Mapping logical connectionsbetween different master data domains and reference data illustrates that both kinds of referencedata (public and private) have a large number of connections to every MDM domain. This meansthat an error in reference data will ripple outwards affecting the quality of the master data in eachdomain, which in turn affects the quality in all dependent transactional systems. The heavilyinterconnected nature of reference data is why it requires separate management and governance.Clearly, Reference Data Management is a major IT initiative being undertaken by a largenumber of market-leading global 5000 enterprises. Both as an IT discipline and a commercialoff-the-shelf software solution, RDM solutions are being brought to market at an increasing pace.Additionally, RDM is a good entry-level project to show success for initial MDM investment whichcan be built on as a data governance model.BOTTOM LINE: Software AG webMethods OneData is a proven multi-domain RDM hub.Due to the highly inter-related nature of reference data, this semantic model-based MDMsolution provides a flexible RDM solution to provide support for an organization's datamodels for both private and public reference data. Additionally, its strong workflow,hierarchy management, and BPM integration address the need for RDM governance. During2012-13, organizations evaluating RDM solutions should review their use cases andconsider Software AG webMethods OneData solution for RDM applications— independentof other pre-existing MDM investments.

The "Field Report" Methodology2012-13 “MDM & Data Governance Road Map”. Part of the deliverables for our client AdvisoryCouncil is an annual set of milestones to serve as a "road map" to help Global 5000 enterprises focusefforts for their own MDM programs. For planning purposes, we thus annually identify ten milestonesthat we then explore, refine and publish via our MDM Alert research newsletter. This set of "strategicplanning assumptions" presents an experience-based view of the key trends and issues facing ITorganizations by highlighting: MDM, Data Governance, Customer Data Integration (CDI), ProductInformation Management (PIM), and (as of 1H2012) Reference Data Management (RDM).Thus the 2012-13 MDM road map helps Global 5000 enterprises (and IT vendors selling into thisspace) utilize these “strategic planning assumptions” to help focus their own road maps on large-scaleand mission-critical MDM projects. During the following year, we use these milestones as the focus forour analyst research in that every research report we write either confirms or evolves one or moremilestones as its premise:1. Pervasive MDM6. Social MDM2. Data governance7. Identity resolution3. Business process hubs8. Big data4. Universal MDM9. Business-critical MDM5. Reference data10. Budgets/skillsAs an industry-funded multi-client study, the MDM Institute is releasing its "Reference Data Management:Market Review & Forecast for 2012-15" during 2H2012. Among other benefits, this industry report providesinsights into: what is RDM, what are the business drivers for RDM, what are the major use cases, what are thetechnical challenges, who are the major solution providers (software vendors and consultancies), how toevaluate such solutions, and what are the best practices for RDM in the large enterprise. Additionally, the MDMInstitute is providing a series of Field Reports that will provide details on the merits and caveats of the variouslymarketed commercial multi-domain RDM solutions. Please da.html for the latest editorial calendar of such Field Reports.The majority of this Field Report on Software AG's webMethods OneData capabilities thereforerepresents our analyst opinion buttressed by in-depth reviews, evaluations and (often) hands-on proofof-concepts executed by the membership of the MDM Institute's Advisory Council.

Evolution of webMethods OneData as a Reference Data Management SolutionSoftware AG's webMethods OneData solution is most often deployed as a multi-domain MDM solution.In our informal surveys, the majority of Software AG's customers also use the software for RDM (inaddition to other domains such as products, customers, assets, and locations). While there have been anumber of solutions that specifically address the straight through processing (STP) and real-time requirementsof the capital market industry (i.e., Asset Control, Eagle, GoldenSource, et al), the wM OneData product isarguably a 4th generation MDM hub (SOA architecture, integrated data governance, integrated BPM) that hasfound frequent uptake as an RDM solution in large enterprises. In short, the Software AG MDM platform alsoprovides out-of-the-box RDM services to centrally create, change, import from external sources (e.g., Dun &Bradstreet), govern and distribute reference master data across an enterprise's entire landscape.Reference data is but one of multiple domains that webMethods OneData has found success in"mastering". In short, it treats reference data as yet another type of master data domain similar to the way theMDM software manages CUSTOMER and PRODUCT master data. This provides a number of majoradvantages to the deploying organization: Common MDM support services to administer and deploy Common MDM infrastructure to manage Full Service-Oriented Architecture (SOA) support with integration to Software AG's BPM solution Ability to connect reference data to both customer and product master data in addition to managinghierarchies across other external data setsIn our informal survey of the customer base, we found Software AG's multi-domain MDM hub used asan RDM solution in the majority of their customer base. We also found that approximately 95% ofwebMethods OneData customers are using the software for more than 2-3 domains. Concurrently, we havefound that many IT departments are willing to consider a standalone RDM solution as an adjunct to their megavendor MDM (e.g., IBM, Informatica, Oracle or SAP). This is because IT departments have found that MDMsolutions that are designed for a specific domain such as Customer (Informatica MDM, IBM MDM Server) orProduct (PIM solutions) are harder to adapt to RDM requirements. Software AG's MDM software can alsomanage the reference data used to map dimension and attributes across analytical tools or reporting systems,especially. The Software AG product has already been tested and vetted as an RDM solution at a number oflarge global enterprises such as Allianz, Avon Products, Chubb Insurance, General Electric, Givaudan, Nokia,Parexel, Pepsico, State of New Jersey, SwissGrid and Wellpoint for both “public” and “private” forms ofreference data.Summary Evaluation - Top 10 Evaluation CriteriaAs part of the interactions with its Customer Advisory Council, the MDM Institute captures and promotesmodels such as "top 10 evaluation criteria" for key MDM-related subsystems. During 2H2012 and as part of thebackground research for the much more comprehensive “Reference Data Management: Market Review &Forecast for 2012-15” report, more than thirty Global 5000 size enterprises shared their software evaluationprocesses and also contributed commentary and supporting details for a set of "top 10" evaluation criteria forRDM solutions. These evaluation criteria (figure 1) are discussed in more detail in the above referenced marketstudy. The majority of this Field Report in turn takes these "top 10" evaluation criteria as a framework todiscuss and understand the capabilities of Software AG's MDM solutions as an RDM Hub.

1. Ability to Map Reference Data — webMethods OneData's MDM capabilities include the ability to maintain acanonical view of reference data to enable the creation of a "standard" across the enterprise. However, notevery application can consume or use the canonical representation. An RDM hub therefore must be able tomanage application-specific or local representations of a reference data set in addition to the canonical dataset. Software AG's built-in Metadata Registry capabilities are used for metadata mapping between themetadata model in OneData and external systems.One can also use the built-in OneData InterchangeMapping functionality which allows for inbound andFigure 1 outbound mappings from/to data definitions, sources andRDM Evaluation Criteriadestinations such as flat files, xml files and databases.1. Ability to map reference dataThese interchange mappings can be accessed either from2. Administration of reference datathe OneData UI or by using OneData’s RESTful webtypesservices.3. Management of reference data sets2. Administration of Reference Data Types — One of4. Architecture/performancethe common problems with custom-built RDM solutions is5. Hierarchy management over sets ofthat the many different types of reference data cannot bereference dataeasily represented by a single data model. The data model6. Connectivityneeds to be constantly changed and extended to support7. Import & exportnew reference data sets, and new properties specific to the8. Versioning supportvaried types of reference data being managed. Thistypically requires substantial development work and IT9. Security & access controlintervention. With webMethods OneData the administration10. E2E lifecycle managementof reference data "types" is the same as for any other dataSource: The MDM Institutetype. OneData is able to categorize the data usingreferences and hierarchical data structures maintained inthe form of conceptual objects — where a "conceptual object" is a logical object, as opposed to a physicalobject. This enables logical groupings of related physical objects and all their attributes for easier (group)management.3. Management of Reference Data Sets — With Software AG's RDM capabilities, management of referencedata sets is the same as any other data set, with User Role management, versioning, security, workflows, datagovernance, etc. By providing intuitive UIs and a flexible data model, webMethods OneData enables anenterprise to quickly install, configure and import reference data with minimal need for ongoing IT involvement.Software AG's MDM hub enables reference data stewards to immediately perform role-based CRUD(create/read/update/delete) operations over an enterprise's reference data sets — with full end-to-end (E2E)lifecycle management and versioning. With the business user as the design point, all of the UIs andstewardship processes are thus defined for RDM, not MDM. This is in contrast to RDM solutions built as acustom domain on a multi-domain MDM platform. Such RDM-via-custom-domain solutions typically entail moreinitial implementation work than a purpose-built RDM packaged offering. In addition, the custom build approachusually requires additional development effort on an ongoing basis. OneData can also manage data sets thatreside outside of the OneData repository using its Remote Object capability. This enables the collectivemanagement of reference data residing in the OneData repository as well as reference data outside of it. Theadvantage is that this makes management of data sets in those objects much easier and is a uniquedifferentiator relative to other commercial RDM solutions. In interviews with RDM consultants and Software AGcustomers, we found that there are sophisticated and large IT organizations that have applied OneData for thispurpose on top of (in addition to, and integrated with) their Informatica or Oracle operational MDM hubs.

4. Architecture/Performance — A critical dimension of RDM platforms, in terms of architecture, isaccessibility. Managed reference data must be made easily available to downstream systems, remotesubscribers, or anyone interested in these data sets. Knowing that each system and each subscriber wouldprefer to access the reference data in a way that is most convenient to them, a flexible RDM solution must beable to expose the reference data in many different ways. This includes real-time channels such as JMS, ondemand access using SOAP or REST web services, on-demand access or scheduled deployment to flat files,xml files, or direct connections to remote databases.Figure 2 Each of those channels should also allow for retrieval ofOverview of Software AG'sall data sets via lookups of specific entries. Aninterdependent factor to architecture is naturallywebMethods OneDataperformance. Making data available to numerousas an RDM Solutionsubscribers creates a need for an in-memory solutionSTRENGTHSthat enables the RDM hub to keep up with high SLA1. Robust solution for centralized datathroughput and latency standards. OneData supportsgovernance, management, stewardship,these accessibility and performance requirements via& distribution of enterprise referenceTerracotta — the de facto standard in Java cachingdatatechnologies — with the option to upgrade to Terracotta12. Proven multi-domain RDMBigMemory for larger scale and higher volumes.5. Hierarchy Management Over Sets of ReferenceData — With OneData, hierarchical structures can beleveled, self- or network-recursive, or have any parentchild type relationship. Software AG provides multipleways to manage such hierarchies including:browsing/navigating via an applet-based hierarchygraphical viewer, linking individual hierarchycomponents together to form a "conceptual object"(e.g., to purge/restore/delete hierarchies), as well asmanaging time variance via "temporal objects" toimplement "effective dates" and "expiration dates" (e.g.,to allow users to query for hierarchies at a “point intime"),3. Deep cross-referencing/hierarchycapabilities with impact analysis lackingin MDM solutions4. Strong taxonomy support/mappings forpublic & private reference data5. Can ingest customer metadata models &UI wraps around model6. Model-driven ease of deployment,implementation, & use (integration withBPM for SOA & MDM deployment))7. Stable, deep pockets vendor2CAVEATS1.Nascent market presence2.Shortage of knowledgable consultancies6. Connectivity — It is vital that an RDM solutionprovide multiple, flexible means of connection to provide 3. SaaS capabilities not marketed4. Under invested in marketingmaximum "accessibility". Reference data must bemade easily available to downstream application(1) Allianz, Avon Products, Chubb Insurance, General Electric,Givaudan, Nokia, Parexel, Pepsico, SwissGrid, Wellpoint, systems, remote subscribers, etc. Further, each(2) 2012 est. revenue 1.2B, 5K employees, 70 countriesconsumer of RDM data must be able to access the dataSource: The MDM Institutein a means and format that is most convenient to them.Therefore, RDM solutions must be able to expose thereference data in multiple, flexible diverse ways. OneData exposes RESTful Web Service links for each of itsobject types (conceptual, temporal, etc.). JDBC, JMS, and file based interfaces are also available for importingand exporting data to and from OneData. Also, via the MDM hubs' "remote object" capability, it can also canmanage data sets residing in other data hubs as noted previously.#3 – Engage all levels of management &adjudicate between centralized vs.decentralized data stewardship#4 – Evolve key stakeholders from “dataownership” to “data stewardship”

7. Import and Export — Software AG's MDM hub provides import and export of reference data in multipleformats, including JMS. For example, for inbound and outbound mappings from/to data definitions, sourcesand destinations such as flat files or databases as well as CSV and XML formats. Wizards guide the userthrough the process of mapping the import columns to the reference data set properties within the hub. Datacan be imported into one object at a time ("data object" import) or into several objects at one time ("conceptualobject" import). Data can also be imported manually through the Data Manager or scheduled via an Import JobManager. The advantage of using the Import Job Manager is that an job can be saved and scheduled to run ata specified period of time. The export mechanisms in OneData use Distribution Jobs that provide a “push”mechanism to deploy data and content to subscribing external systems. Distribution Jobs can use multipledelivery mechanisms such as remote database updates, message queues, or files with FTP transfer options.Alternatively, any ETL tool can be used seamlessly with OneData. Reference data can be also “pulled” fromthe central repository in OneData using its generic API and Web Services functions.8. Versioning Support — Versioning is required to effect end-to-end lifecycle management to managechanges to the reference data sets and mappings over time. With OneData's RDM approach, hierarchies arerelated to the version of the set that they are created against. This versioning support manages the lifecycle ofa canonical set, the lifecycle of application-specific or local sets mapped to the canonical, and the lifecycle ofthe mappings themselves. Hierarchy management capabilities in OneData include functionality formaintenance, versioning, approval and release of hierarchy versions. This enables OneData business datastewards or SMEs to manage their data hierarchies — e.g., reference data, product bill of materials, salesterritory structures, customer organization structures, business units, etc. — as “one version of truth” aboutthese key enterprise business relationships. OneData also allows versions to be created of changes made for“what if” considerations, for analysis or comparison activities. Yet another versioning option is ArchiveSnapshot which enables creating a version of a particular data object related to its “as of” date and assigningthe object as a temporal object -- resulting in each entry having an effective date and expiration date. Thisenables users to browse the object from an “as of a specific date” to obtain a historical view of the data.9. Security and Access Control — Software AG's RDM solution provides robust role-based security. Forexample, with OneData CRUD access to a particular entity is controlled by the user’s role, the group that theuser is in, and related ownership of the entity, and the life-cycle state of the entity itself. This role-basedauthorization is configurable and integration with LDAP is supported. The OneData open data model enablesthe building and maintenance of custom categorizations via approval workflows and authorization schemes.Data authorization in OneData can be very granular — from object level to particular object attribute level(vertical) or based on data values itself (horizontal).10. E2E Lifecycle Management — OneData supports formal governance of reference data, putting end-to-endlifecycle management of enterprise reference data in the hands of business users — reducing the burden onIT, and improving the overall quality of data used across the organization. The user interface is a role-based UIwith built-in security, versioning, and review and approval lifecycle management. OneData's workflow moduleis a "state machine" which manages the end-to-end lifecycle of reference data. The workflow module facilitatesa follow-up approval process once changes to values have been made. All data changes may be staged androuted through an automated approval process to ensure data governance policies are in effect.

Competitive OutlookCompetition for a multi-domain RDM product such as Software AG's includes: Custom-built, manual solutions Multi-domain RDM Hierarchy management system adaptations Purpose-built or industry-specific RDM Custom MDM domain typeCustom-Built, Manual Solutions — Many enterprises struggle with home-grown RDM using spreadsheetsand other error-prone manual processes to manage to reference data sets and their relationships to each other.Just as customer-built CRM, ERP and MDM etc. have faded when commercial off-the-shelf solutions becamewidely available, so too will manual RDM solutions fall into disfavor. With custom-built or home-grown RDMsolutions stewards have to rely on IT for changes to functionality and are unable to change the business rulesrelating to the reference data themselves.Hierarchy Management System Adaptations — Organizations can attempt to use simple hierarchymanagement software, but such systems do not readily support publish-subscribe, classification mapping, etc.Examples include: Microsoft Master Data Services (MDS) or Oracle Hyperion Data Relationship Management(DRM). Many finance departments use tools such as Oracle DRM for financial hierarchies and attempt to applythese tools to hierarchies in human resource assets, location assets, etc. This approach has not provenenterprise-scalable in our experience. Because webMethods OneData is a multi-domain MDM hub, it ispossible for it to handle both hierarchies for finance departments and other domains, as evidenced by theircustomers who are doing so. We believe Software AG is working on other add-ons to target hierarchymanagement in other solutions.Custom MDM Domain Type — Both Informatica (Informatica MDM) and SAP (SAP NetWeaver MDM and SAPMaster Data Governance CUSTOMER object) offer the capability for custom domains to be created andmanaged to implement reference data management. Reports from organizations that have gone this routeindicate that it is not as easy to implement RDM as a custom domain type as these vendors promote. In multidomain MDM solutions originally designed for managing customer data (e.g., Informatica MDM, formerlySiperian), organizations report lack of data modeling flexibility, rudimentary lifecycle management capabilitiesand limited data governance features, in particular around authoring, workflow and cross-temporal relationshipmanagement.Multi-Domain RDM — Certain of the commercially available MDM products were architected with semanticlayers on relational DBMS which provided flexibility in defining and managing multiple domain types (hence thename "multi-domain" or "multi-entity" MDM). While these products provide good flexibility and ease of use, themarket feedback is that certain of these systems incur substantial processing overhead when attempting toscale into a large-scale enterprise solution. As an architectural/performance tuning option, certain of theseRDM solutions (such as webMethods OneData ) offer the ability to run parts or all of the RDM/MDM platform ona standard relational DBMS platform.Purpose-Built or Industry-Specific RDM — Certain enterprises have used SAP's PIM solution as aconsolidation type of RDM support. For example, consider SAP's "item master" with its staging areas and minimodel for landing reference data which also includes simple workflows. There are also purpose-built RDMsolutions which leverage the hierarchy management capabilities of a mainstream MDM platform such as OracleMDM or IBM MDM — Oracle's Site Hub and Kingland Systems' Security Master are examples. Otherorganizations have attempted to manage look-up tables such as RDM data via an existing AssetControl, Eagleor GoldenSource real-time RDM by simplifying what features are used. The challenge in this scenario is thatthese premium priced real-time RDM solutions do not represent good economic sense.

Futures for Software AG's MDM Hub as an RDM SolutionIt is our view that Software AG will continue to have success in large enterprises, continuing its momentum inthe consumer packaged goods (CPG) and financial services industries. Software AG is also well position toaddress small-to-,medium businesses (e.g., US 500M to US 1BM revenues) and departments of very largeenterprises (for example, RDM for finance & accounting, RDM for human resources). Software AG has astrong vision and road map for RDM going forward given its investments in coordinating its BPM and MDMportfolios. Some of the key areas we believe Software AG should focus on include: improving integration with3rd party business process management (BPM), adding enterprise content management (ECM) support forunstructured information, and providing a Cloud-based (Software-as-a-Service) RDM capability.Currently, the webMethods OneData solution is focused on multi-domain reference data and other MDMdomains that require strong governance due to complex cross-domain entity relationships (and the hierarchiesentailed). In the future, there should be more functionality in supporting broader scope operational MDM usecases such as pre-packaged B2C customer data integration (CDI) and product information management (PIM)use cases. Organizations are also asking for improved impact analysis to understand the impact of an RDMchange in one table to downstream consuming systems, as well as data lineage. Organizations also areincreasingly asking for graph-styl

solutions that are designed for a specific domain such as Customer (Informatica MDM, IBM MDM Server) or Product (PIM solutions) are harder to adapt to RDM requirements. Software AG's MDM software can also manage the reference data used to map dimension and attributes