Professional Documents
Culture Documents
With all of these unsettling forces, little attention has been paid to the best way to manage the
correlation of these new data structures into integrated business and technical descriptions. These
connections between technical and business domains are critical to pull various data structures together
into operational systems or analytical platforms and share them across platforms and business units in
a timely and efficient fashion. Correlation between various data structures and definitions also enables
additional data management activities such as data governance and data quality that lead to improved
confidence of those integrated business views of a companys data.
Business Agility
Speed, flexibility and dexterity are all components of a concept that many refer to as business agility
the facility to react quickly to changes in business climate as well as evolving technical developments.
Business Agility enables organizations to build or extend their competitive edge over their rivals. The
proliferation of disruptive new data sources and the use of siloed design techniques for managing the
information about data sources (e.g. metadata) will quickly hamstring any organization from acting
nimbly. That is unless they have a robust data management process in place that can account for these
new forces and turn them into a competitive advantage.
Often with the advent of new technologies, more traditional documentation is seen as unnecessary
overhead. On the contrary, a set of enterprise-wide definitions of business concepts such as Customer,
Product, Employee, Revenue and Loss is more important than ever. Improving how existing data source
metadata is documented and new sources are integrated allows both business and IT stakeholders to
have the visibility and access required to make rapid decisions. A new way of managing business and
technical definitions, existing and new data models, and sharing that metadata information throughout
an organization is essential to this process.
If new data platform technical descriptions are mapped to centralized business definitions as each one
is introduced into an existing environment, time to implementation for those systems is reduced. With
maps to centralized business definitions, modification to downstream applications such as analytical
reporting platforms and operational systems is easier. This concept is based on having a single set of
definitions and in a single repository. Establishment of a coordinated environment for the management
of data models and metadata facilitates Business Agility in an organization.
Page 1
In the early days of data management maturity, projects were implemented using siloed development
paradigms internally by IT teams. Rudimentary data models based on technical metadata and
department specific business definitions were documented with standalone methodologies using siloed
documentation products or database specific design tools that locked information on a single laptop or
a single database server. This paradigm was well designed in terms of technical database design, but not
much else. Inputs in terms of data models and other metadata were based on technical platform type.
Outputs in terms of documentation and coordination were manual based on one-on-one meetings and
emailing of documents.
As data management practices matured, centralized data modeling repositories were developed.
Technical metadata from one or more database platforms was integrated with business definitions.
Business stakeholders had access to technical details of model metadata. Members of IT teams had
visibility into business metadata across departments. Both teams typically were required to access
these environments via desktop application. Also, centralized repository users often needed specialized
knowledge to understand both technical and business aspects. These new centralized repositories
dramatically improved the correlation of metadata inputs into a single environment. Communication
outputs were still a manual and relatively asynchronous affair.
As we approach an era of expanded requirements on data management practices brought on by
disruptive, non-relational datasets, the opportunity for a truly coordinated data management
environment for metadata and data models is now available. This
environment capitalizes on the development and maturity of collaborative
tools to bring bi-directional communication between business and
The opportunity for a
technical stakeholders and close the loop between these two groups. A
truly coordinated data
Coordinated Data Management Environment (CDME) automatically
management environment
maintains links between business and technical metadata definitions
for metadata and data
stored within a modern data model. It is a nimble repository based on
models is now available.
the following information:
Description of business definitions common across departments.
Definition of structured/relational data models, whether they are operational systems (e.g. customer
care, point of sale or billing, analytical platforms (e.g. data visualization, reporting or advanced
analytics) or data integration tools (e.g. extract, transform and load ETL).
Mapping of new datasets from relational (e.g. acquired inventory control system or new customer
relationship management platform) or multi-structured sources (e.g. sensor log repository or social
media environment) to the existing definitions in operational or analytical systems.
Collaborative communication layer to coordinate efforts between disparate stakeholders.
Flexible output formats for search and export to enable distribution of information and designs.
This coordination supports the type of Business Agility required in modern organizations. As with
centralized repositories, business teams input the semantic designations. IT teams input technical
definitions. However, both business and IT can see changes of the other because their independent
work is guided by the CDME and presented as appropriate to the reader. Business stakeholders
look at the metadata across multiple systems and platform types to dynamically assist with integration
designs and implementation choices. Technical stakeholders can see the semantic definitions of such
core concepts as Customer, Product and Employee across multiple departments. From this view,
determinations can be made on how new data sources and platforms can complement existing data
models, or require net-new definitions.
2
Page 2
A decade ago, there were a limited number of data management platforms. Today, there is a wider set
of options. A Coordinated Data Management Environment is critical to the enablement of Business
Agility of this larger environment. Operational Systems, whether based on mainframe data stores or
relational databases, were on one side. Analytical platforms based on Data Warehouse or Data Mart
structure were on the other side. Recently, additional options have emerged. Analytical Appliances
now speed the processing and implementation of analytical processing. Cloud-based Platforms speed
time to implementation. Hadoop and other NoSQL Platforms emerged because there was a need to
process the datasets that traditional platforms could not process or were prohibitively priced to do so.
Finally, Data Discovery platforms now bridge the gap between the SQL data platforms and NoSQL
platforms. Enterprise Management Associates (EMA) has defined this collection of platform choices as
the Hybrid Data Ecosystem (HDE).
Line of Business
Executives
AL ANALYT
RATION
ICS
OPE
Business
Analysts
BI
Analysts
TION MANAGE
ME
RMA
NT
FO
N
I
Discovery
Platform
ANALYTICS
LOAD
STRUCTURE
ECONOMICS
REQUIREMENTS
COMPLEX
WORKLOAD
RESPONSE
Analytical
Platform
(ADBMS)
ONAL PROCES
SIN
ATI
ER
G
OP
Enterprise Data
Warehouse (EDW)
Cloud Data
DA
Data
Scientists
Hadoop
N
TA IN
T EG R A T I O
SQL
External
Users
Operational
Systems
NoSQL
EXPLORATION
Developers
IT Analysts
Page 3
24%
22%
Compound (XML)
18%
18%
16%
2%
0%
5%
10%
15%
20%
25%
30%
The preference for structured data schemas comes from the fact that to be effective many business
challenges still require a relational model. These challenges include descriptive analytics associated with
standard reporting, and normalized models connected to operational systems. Over 60% of the HDE
platforms implement some form of traditional structured data schema. Operational Systems, Data
Warehouses, Data Marts, Analytical Appliances, and many Cloud-Based Platforms implement
on structured datasets based on Boyce-Codd Normal forms popular with operational platforms or
denormalized structures important to analytical platforms. It is important for organizations to integrate
and map new data sources to traditional relational data schemas that are central to the organizations
processes and applications.
With new HDE platform types comes the need to manage a new set of data structures. Often these
structures are referred to as unstructured because there usually is a human-generated content element
within free form areas. Often these unstructured areas are contained within a larger structure on an
application or other business document and should be characterized as multi-structured. Multistructured data formats can have issues with access methods such as standard query languages like SQL.
Page 4
Because no one platform can meet all the challenges inherent in the
current state of data management, these platforms are used in concert.
Many times the information in a NoSQL platform can be modeled and
transformed into a standardized SQL environment. And since, as we
saw in Figure 2, most organizations have the majority of their data in
relational systems, it is critical that these new types of data sources can
be integrated with existing system definitions. With a Coordinated Data
Management Environment, this process becomes an incremental effort as opposed to a series of oneoff projects with repeated levels of effort, and there is no need to start from scratch and reinvent the
wheel with each new project core data definitions can be reused and leveraged in new applications.
From analytics and reporting to database administration to new areas of data-centric application
development, the ability to effectively manage metadata and provide visibility into data model structures
will be the key allowing these disparate groups to manage data across initiatives.
With a CDME, the following areas will be able to re-use common models and definitions without
building from scratch:
Advanced analytics
Application development
Business intelligence reporting
Enterprise architecture
Database administration
Cloud-based infrastructure
Master data management
Data governance and data stewardship
The Business Agility, enabled within these areas, provides the foundation of proactive data management.
This includes reduced time to implementation for projects such as integrating data from cloud-based
platforms with in-house applications; reduced costs relating to rework between existing data models
and new model development; and improved levels of analytics and reporting across the organization.
Page 5
Ideally, data governance is a cross-team effort. IT teams understand the underlying technical
implementation aspects such as data types, constraints and loads. Business stakeholders are responsible
for the semantic definition of customer, product and employee. It would only seem natural that if you
were going to have an effective data governance program, it would include participation from business
and technology.
Yet more often than not, data governance is assigned to the group that has the most access to the data
IT. This puts IT teams that may not have a true grasp of how business stakeholders use the data in the
drivers seat. A CDME with collaboration tools allows for automated coordination, breaking down the
walls between the two groups and turning data governance into that ideal joint effort.
By linking semantic business definitions with technical implementation information, a central facility
allows for the collaboration between business and IT to be much more efficient. Each group assesses
information in the format and view that best influences its role(s) in the organization. A technology
stakeholder can coordinate with representatives from the line of business without needing to translate
from technical details to business semantics and back again. Being able to actively collaborate between
the groups with a translation mechanism enables the iterative nature of Business Agility.
By including various business stakeholders in the process when adjustments are made to data definition
and management, the door opens to involve business stakeholders in the data quality assurance aspects
of data governance. If communication is intuitive, business stakeholders can use the collaborative tools
to make requests for adjustments when they see data quality issues in their semantic definitions.
To Improve is to Change
Winston Churchill once said:
Page 6
EMA Perspective
Implementation of
a Coordinated Data
Management Environment
makes faster response,
and thus Business
Agility, possible.
Using this type of facility, organizations can match speed of implementation and coordinate between
the myriad new sources and data formats at the speed of data delivery required for Business Agility.
Organizations, which develop required speed of implementation to meet the challenges presented by
the disruptive forces in data management, will have an advantage over their peers in the marketplace.
If organizations, in particular IT departments, are not responsive to business change, business
stakeholders will force change. For example, the Chief Marketing Officer (CMO) and the Marketing
organization may push to have its own operational and analytical environments either in the corporate
data center or in the cloud.1 Cloud-based systems are breaking down corporate firewalls to offer
solutions that the line of business can utilize without direct control by the CIO. All of this expands
the number of platforms that contain data that can and should be utilized by an organization to either
develop competitive advantage with revenue creation or cost reductionor both.
Enabling Business Agility is not just an activity for internal technical and business teams. Software
product vendors are recognizing the need for Business Agility from their clients. CA ERwin, a leading
vendor for data modeling, has recognized this evolution in the market. New data sources and ways
of collaboration with internal and external stakeholders drive the requirement to integrate between
platforms and tools. CA ERwins enhancements have built upon their traditional strengths in data
model and metadata management and have moved to establish a CDME framework where business
and technical stakeholders can work together on integrated views of a companys information.
To create a Coordinated Data Management Environment, vendors must work together to ensure
metadata interchange. EMA has seen that in conjunction with CA Technologies efforts, other software
vendors are recognizing the advantages of coordinated data management in their applications, allowing
an even broader set of users to tap into a shared business and technical metadata. Casewise, a provider
of business process management and business architecture solutions, has integrated data modeling by
OEM-ing the CA ERwin package with its offerings.2 For analytical platforms, BIReady is utilizing a
1
H
artman, Glen, How To Close The CMO-CIO Trust Gap And Bolster Corporate Growth, Forbes.com, August 28, 2013,
http://www.forbes.com/sites/onmarketing/2013/08/28/how-to-close-the-cmo-cio-trust-gap-and-bolster-corporate-growth
Casewise Signs OEM Agreement with CA Technologies to Help Businesses Ease DecisionMaking, Casewise.com, April 29 2013 http://www.casewise.com/news-and-events/news/
casewise-signs-oem-agreement-with-ca-technologies-to-help-businesses-ease-decision-making
Page 7
Page 8