Semantic MDM systems design concept

Semantic MDM systems design concept
“Evolution of society is connected in particular with development of
communication means of its members, especially tools for building and using their
collective memory”

Stanislav Yankovskiy

“A properly organized Semantic Network can facilitate evolution of all human
knowledge at large”

Sir Tim Berners-Lee

Andrey Andrichenko,
Ph.D. in Technical Sciences,
CEO SDI Research
andrichenko@sdi-solution.ru
www.sdi-solution.ru

The design automation systems have drawn closer to the threshold, beyond
which the snowballing use of semantic technologies follows. The interest to these
technologies is shown everywhere, where there are complex data structures and
difficult to formalize decision-making procedures operate, which are based on
empiric knowledge about behavior and interaction of objects. The use of semantic
data models in CAD, CAM, CAPP shall allow creating a new class of intellectual
systems with powerful decision-making.
All objects in production are in continuous interaction viz. materials,
component parts, equipment, jigs, fixtures and tools. The characteristics of these
objects are stored in separate databases, and rules of their behavior and compatibility
– in the algorithms of various applications. By uniting data and knowledge into a
single semantic model of the application domain, intellectual information system of
the enterprise can be built, which shall serve as the basis for taking adequate
decisions in design, production and management.
By definition, the semantic network is an “information model of the
application domain, having oriented graph appearance, whose peaks correspond to
the application domain objects, and directed lines (edges) give the relation between
them” (fig. 1).

1

Fig.1. Fragment of semantic network

The evolutionary development of software consists in gradual unification of
system-wide components. In the next 5 years, the shift of accent from software
development towards generation of application-oriented semantic data models is
inevitable. The standardization and unification of terms, concepts and relations, used
in these models, shall become the key factor in developing any information system.
The replacement of object paradigm by semantic and unification of data models is the
mainstream, which shall allow increasing the automation level of decision-making
and standardize information exchange protocols between different applications.

Fig.2.Unification of system-wide components in the course of evolution of software
tools

2

Historically, the appearance of a new class of systems, designed for
implementation of data domain semantic models, is inevitable. The applications of
Master Data Management (MDM) class, consolidating all enterprise reference data of
non-transactional nature, can serve as the beneficial environment for building these
models.
Under this direction, the problems of duplication and synchronization of
reference data are eliminated. A single classification and coding system is introduced.
A centralized system of storage, management and access to reference data is
implemented, and the prospects for standardizing data presentation and exchange
appear. The “venue” is opened for deploying mechanisms that are based on
knowledge.
MDM methodology analyzes reference data, circulating in the enterprise, as a
single communication language of corporate information systems. It is understood
that information about the products is subject to joint use and exchange only if the
same reference data is used by both the sender and the recipient.
Thus, we deal with innovations in the field of reference data consolidation,
unification of their processing services, and consolidation of knowledge in semantic
models and standardization of data exchange formats.
The development perspective of MDM systems is to absorb the innovations
listed, and together with DBMS class applications, become system-wide IT
infrastructure components of any enterprise.
Let us examine the basic principles of building semantic MDM systems.
Data consolidation
The reference data repository shall be the only place, where data addition,
change or deletion shall take place (fig. 3). MDM is an independent class of systems,
which shall not occupy a subordinate position in relation to any application system,
for example, ERP or PDM.

Fig.3 Reference data consolidation

3

Consolidation of knowledge
The shift of decision-making rules to data model level makes them available to
all corporate applications. The orientation for building semantic models of data
domains provides maximum level of automation, as the isolated solutions, once input
to the semantic reference data database, shall duly be formalized and used repeated
number of times in various application systems (fig.4).

Fig.4. Consolidation of information

Common information space
The semantic MDM system represents consolidated reference data space.
Information is gathered from the primary systems and integrated into the common
permanent storage area. The removal of part of the references beyond its limits
breaks the connection between objects that violates the integrity information system
of knowledge and considerably limits the possibility of building a semantic network
(fig.5).

Fig.5. Common information space of reference data
4

Universality and expandability
The data domain model is regularly corrected and improved. New objects are
created, their rules of conduct and relation change. The semantic MDM shall be
capable of adapting to these changes i.e. in essence, be the medium of performance of
data domain model, independent of its specific content.
Context-sensitive data display
MDM system shall provide the opportunity to “see” objects from different
points of view. For example, the technologist must see the workpiece and cutting tool
displacement mechanism in the metal cutting machine tool, and the mechanical
engineer – units and parts, subject to maintenance inspection (fig.6).

Fig.6 Contextual point of view on reference data information object

The contextual point of view on the object is not limited only the user’s role, it
changes depending on time, to be precise from the stages of object’s life cycle, and
set of its functions (intended use).
The tangible objects have two main properties: structure and activity. The
contextual representation of the object’s internal structure changes dynamically
depending on the processes, in which it takes part. It can be said that the objects are
defined by candidate actions with them.

Standardization of data exchange formats
The topic of data synchronization and unification is far outside the interests of
establishments. In accordance with the requirements of international standards, the
product suppliers shall provide the buyer with technical information about the goods
required for cataloguing in electronic format. The integration of goods from different
manufacturers in the electronic catalogs implies that when describing the goods it is
required to use the same dictionary terms and notations.
As of today, there are two alternatives of data exchange standardization formats.
The first is implemented by ISO 22745 standard, which assumes the use of open

5

technical data dictionary of the International Association for Electronic Commerce
Management (eOTD ECCMA).
eOTD dictionaries are developed with the purpose of connecting terms and
definitions with similar semantic content. They allow to assign unique worldwide
identifier to any term, property or class. On the basis of these identifiers the
description of material and technical objects in various automated systems can be
conformed (fig.7).

Fig. 7 ECCMA Open Technical Dictionary (eOTD)

In conformity with the order of Rostechregulirovaniye No. 1921 of July 19,
2006, the Russian version of open technical dictionary eOTD ECCMA has been
formed that is designed to coordinate information about products from different
suppliers with the purpose of reducing costs for developing electronic catalogs of the
products.
The second choice is realized by ISO 15926 standard, which, in contrast to ISO
22745, is ontological, as it standardizes the structure of objects. The data model,
defining value of information about life cycle in the single context, supporting all
description groups, which the process engineers, equipment engineers, operators,
maintenance engineers and other specialists can hold in relation to the equipments, is
specified in it (ISO 15926, part 1).
The standard data model, based on which synchronization with application data
models is proposed, is implemented in ISO 15926 by reference data libraries RDL
(Reference Data Libraries).

6

Fig.8 Standard model formalizes data exchange

The integration of a new application in the single information space of an
enterprise shall begin with the coordination of classes and attributes of the application
model of this application with the corresponding definitions of the standard model,
which is the corporate communication language of different automated systems at the
enterprise (fig.8).
The real works for using ISO 15926 are conducted actively by the State
Corporation Rosatom and FGUP Sudoexport. On December 16, 2008, an order was
issued under number 710, prescribing: “State Corporation “Rosatom” and its
organizations when forming and using production information models at all stages of
NPP lifecycle and fuel production during information management process for the
purposes of data integration be governed by the provisions of international standard
ISO 15926, for which develop the corresponding corporate standards”.
Semantic technologies in CAPP
Computer aided process planning (CAPP) systems, operating at machine
building enterprises, are the principal consumers of reference information. Data about
material and technical objects viz. equipment, materials, machine tool attachments,
are required by them with maximum details. The interest for CAPP is represented not
only by the technical parameters of objects, but relations between them in the context
of production process. The possibilities of semantic MDM systems allow CAPP
applications to realize “intelligent” search in reference data database, where
parameters of the unknown object, as well as rules of its interaction with other
objects, take part.

7

Thus, for example, when searching for a cutting tool, as criteria one can specify
not only its characteristics, but any other object interconnected with it: material of
machined part, machining scheme, fixtures, metal cutting machine tool. The system
shall select the required tool, compatible with the copies of associated objects (fig.9).

Fig.9. Narrowing of search area in the semantic network of interconnected
objects
Semantic search is a key consumption value, capable of providing competitive
advantage of CAPP due to increase in automation level of decision-making during
the design process.
This approach is the basis of Semantic Web technologies. Semantic
technologies have already passed the initial development stage and are seriously
considered by leading analysts as real force: “During the next 10 years, Web-based
technologies will improve the ability to embed semantic structures in documents, and
create structured vocabularies and ontologies to define terms, concepts and
relationships…” Analytical report “Finding and Exploiting Value in Semantic
Technologies on the Web”, Gartner, 2007
As defined by Thomas Grubber, ontology is the specification of some data
domain, which describes multitude of terms, concepts and classes of objects, and
interactions between them. Ontology is designed to provide coordinated unified
dictionary of terms for interaction of different corporate information systems.
The simplest example of building ontology is to segregate the connecting and
cutting part in the rotary cutting tool structure as independently classified objects that
allow using them when building descriptions of similar tools, like: drill, broach bit,
reamer bit, end milling cutter etc. (fig.10).

8

Fig.10 Allocation of component parts of a rotary cutting tool

Without building an ontological model of the object, it is impossible to
formalize its interconnection with other entities, since the compatibility rules of two
objects are defined by combined compatibility of their component parts (fig.11).

Fig.11 Compatibility of objects is defined by complex compatibility of their
component parts

Integration of information on unified descriptions of application domain
objects in the common library and provision of access to it from different applications

9

resolves the issue of data exchange format standardization. The placement of such
library on the Global Network resolves the problems of data integration at industrial,
state and inter-state level.
Under the European project JORD (Joint Operational Reference Data),
starting 2008, a library of ontological data models has been created based on open
international standard ISO 15926. Each volunteer has the opportunity to place his
own ontological data models in this library. The annual subscription for this library
in the internet shall cost EUR 25,000.
Semantic, corporate reference data management system
SDI Solution informs about the release of Semantic, a new corporate reference
data management system. This software solution has advanced functional of
information and search system and simultaneously serves as the reference data
provider for CAPP, PLM and ERP.

Fig. 12 Semantic, corporate reference data management system

The Semantic system supports corporate reference data management business
processes: data input, actualization, access, and control, including maintaining history
of changes and data use. It implements multi-criteria parametric and semantic search
of objects. It allows storing data in various media: Oracle, MS SQL Server,
InterBase. Detailed description of functionality of Semantic system shall be presented
in the next issue of “CAD and Graphics”.

10

Semantic MDM systems design concept

More Related Content

What's hot (19)

Viewers also liked (6)

Similar to Semantic MDM systems design concept (20)

Recently uploaded (20)

Semantic MDM systems design concept