设为首页 | 添加收藏
首 页 走进石竹 石竹动态 产品中心 解决方案 服务体系 知识中心 招贤纳士
 
K 知识中心  
数据治理
元数据
数据质量
技术讲坛
下载
 
C 联系方式  
北京:(8610)58810139
广州:(8620)61246699
上海:(8621)62351222
销售联系:
info@carnation.com.cn
技术支持:
support@carnation.com.cn
 
 
  当前位置:首页 >> 知识中心 >> Mapping the Enterprise Genome
Mapping the Enterprise Genome
发布日期: 2017-03    阅读次数: 7480

Robust metadata representation standards emerge to tackle information integration troubles

by Doug Laney


Data warehouse and other developers have long awaited common metadata management and interchange mechanisms broadly supported by vendors of information supply-chain (ISC) development tools. Such mechanisms would increase productivity by making definitional metadata more reusable and available.

As ISC efforts evolve from departmentally instantiated business solutions to integrated, enterprise-level initiatives supporting a new breed of mission-critical business solutions (such as customer relationship management, business process management, enterprise resource management/planning, and supply-chain management), definitional consistency has become a requisite for their success.

In the e-commerce world, effective metadata interchange will be necessary to communicate information using the preferred proprietary hierarchies and categories of each enterprise’s partners, suppliers, and customers. Moreover, by year’s end the reach of these solutions will extend to as many as 10 times the number of intra-enterprise users compared to last year. (See Meta Group’s “1999 Data Warehouse Marketing Trends/Opportunities” industry study.) After that, they will reach extra-enterprise users (customers, suppliers, and partners) as well. This new diverse assemblage of end users will require sophisticated means to assist them in navigating information sources and customizing content delivery.

Even by 2001, a single metamodel standard will still not be widely accepted nor implemented, but batch metadata interchange mechanisms will abound that broker proprietary format metadata among application development and end-user tools. These interchange facilities, based on XML, will evolve to realtime interoperability in 2002 or ’03 to supplant most batch interchange interfaces.

For several years, developers and vendors have glumly watched the two major metadata standards bodies, the Meta Data Coalition and the Object Management Group, jockey to define and promote metamodel and metadata interchange standards. (See Table 1, for a comparison of current standards.) Founding and participating organizations, including extract/transform/load/manage (ETLM) tool vendors, OLAP tool vendors, and major players such as IBM, Microsoft, and Oracle, have strongly promoted their own ideas, many unbefitting to an acceptable industrywide standard for enterprise metadata management.

TABLE 1 “DNA” framework for metadata.

Current Standards

Governing Standards Body

Leading Vendors

Notes

OIM

MDC

Microsoft, Evolutionary Technologies International Inc., and Unisys

·  Covers analysis/design, objects/components, databases/data warehousing, business engineering, and knowledge management

·  Includes draft XML interchange spec Represented in UML

·  Incorporates MDIS v1.0, but not v1.1

·  Broader specification than CWM

CWM

OMG

IBM, Oracle, and Unisys

·  Covers all aspects of data warehouse metadata management

·  Includes ratified XML interchange spec (XMI)

·  Represented in UML

·  Stored in Meta Object Facility (MOF)

·  Broader commitments for implementation than OIM

CDIF

EIA

 

An older format for exchanging definitional metadata

MDIS

OMG

 

CWM usurped this, OMG¹s original metadata specification



The recent Object Management Group ratification of the Common Warehouse Model (CWM) specification, however, represents a major step toward improved data, application, and process integration. Enterprises struggling to capture and synchronize data definitions, simplify inter-application data mobility, and track information as it flows through their information supply chain eventually will benefit from metadata standards such as CWM. Through mid-2001, exclusive metadata interchange standards in the data warehouse and analytical application domain (Informatica’s MX2, Informix’s MetaStage, and the Microstrategy-NetGenesis API, for example) will dissipate in favor of CWM or its opposing Open Information Model (OIM) standard from the Meta Data Coalition. During this time, even the Object Management Group’s contributing vendors will struggle to fully implement CWM, instead offering only XMI (CWM’s XML data type definition) interchange product features and “vendor enablement kits,” including libraries of Java interfaces that abstract away direct XMI coding.

By 2002, the first fully CWM-based technologies will appear, but customers should remain indifferent to the underlying schema of products as long as the products support accepted metadata interchange standards. (See Meta Group Research Note ADS835, “Common Data Models Are for Commoners.”) By 2004, robust federated metadata networks will emerge that provide ubiquitous metadata interchange among information supply-chain tools and include publish/subscribe along with traditional repository services, such as impact analysis, devoid of the need for a central metadata repository. (See ADS825, “Metadata Maelstrom Sinks Standards and Repositories.”)

Metamodel Scope and Support

Unlike MDC’s most recent specification released in 1999, CWM is the result of a pragmatic collaborative effort among OMG contributors (including Oracle, IBM, Hyperion, Unysis, and NCR) and a significant cluster of private industries. CWM leverages parts of OIM, although supporters position the two standards distinctly, and both are specified in the Unified Modeling Language. Also, each specifies divergent metadata interchange XML data type definitions.

IT professionals must remember that the CWM scope is limited to metadata relevant to data warehousing and analytic applications, while the OIM schema is declared capable of handling knowledge management and business-process constructs. Therefore, enterprises considering panoramic metadata and repository initiatives may find CWM limiting, though more broadly supported. IT shops tending toward one or more of the CWM collaborators may find solace in the ability to create Extended CWM (CWMX) model extensions. Microsoft shops, on the other hand, will find comfort in the OIM specification because Microsoft contributed the original version (which subsequently has been retooled and stripped of platform-specific constructs), and continues to support it via Microsoft Repository.

In evolving the metadata specifications through 2002, the standards bodies will include test benches, certification criteria (for both schema and interchange implementation), certification services, and intersections with existing scheduling and transport services that would enable federated metadata networks.

Revamping the Value of Metadata

Metadata management has long been advocated as a critical component of the data warehouse infrastructure, and more recently been branded as beneficial to analytic application users. Meta Group’s recent “Data Warehouse Scorecard and Cost of Ownership” industry study quantified the enormous effect metadata has on the overall success of data warehouse initiatives. (Seventy-five percent of highly successful data warehouse initiatives include formal metadata management facilities.) Still, in the heat of a typical data warehouse battle, rarely is metadata seen on the front line. If the tired, old “card catalog” analogy fails to get management enthused about the value of metadata, consider the notion that metadata is an enterprise’s digital “DNA.” (See Figure 1.) Metadata specifications such as CWM and OIM are tools for expressing the “enterprise genome” and mapping the relationships among all data, applications, processes, and resources. Such mappings help enterprises understand their infrastructure, architecture, and operations, letting them mutate or be spliced (with business partners or merging businesses) for improved business performance.

FIGURE 1 Federated metadata vs. a repository. Enlist data warehousing tools that capture, broker, and publish metadata via XML.


Erstwhile metadata nomenclature includes classifications such as technical, business, and operational metadata. These classes, however, fail to speak to the value of metadata. (See Table 2.) In line with the digital DNA theme, Meta Group’s research finds that comprehensive metadata solutions comprise:

TABLE 2 Data warehouse metadata standards comparison. Robust, integrated metadata solutions will aid in using, developing, and operating the data warehouse.

Metadata: the data warehousing environment's DNA

Metadata Type

Example

Benefit

User

Source

Definitional

Business meaning,

Recognition,

Casual,

Heads, documents,

 

calculations,

understanding,

power,

spreadsheets,

 

system of record

trust

new

process design

Navigational

Alias,

Location,

New or casual,

Data model,

 

canned report,

expedience,

executive, or

BPM/OLAP tool,

 

refresh date

accuracy

frequent

job log

Administrative

Usage,

Performance,

Operator,

Monitor,

 

key attributes,

integrity,

modeler,

data model,

 

mappings

scalability

designer

ETLM tool

© 2000 META Group Inc., Stamford, CT, (203) 973-6700, metagroup.com



Definitional metadata that expresses the meaning of enterprise objects (whether they are data, applications, processes, or resources) to ensure their consistent usage and interchange throughout an enterprise and its evolution. Definitional metadata fosters improved understanding of information sources.

Navigational metadata that provides links and expresses relationships between enterprise objects to ensure their accessibility and ease of use. Navigational metadata helps locate information as the information supply chain evolves.

Administrative metadata<SPAN lang=EN-US style="FON

 
 
关闭窗口
 
石竹软件 版权所有 粤ICP备13017351号 | 公司月刊 | 招贤纳士 | 石竹质量 | 联系我们 | 邮箱登录 |
石竹北京 电话:(8610)58810139
石竹广州 电话:(8620)61246699
石竹上海 电话:(8621)62351222
销售联系:info@carnation.com.cn