Corporate Information Factory
& glossary
Glossary of Data Warehousing
- the operation of seeking, reading, or writing data on a storage unit.
method - a technique used to transfer a physical record from or to a mass
storage device
mode - a technique in which a specific logical record is obtained from or
placed onto a file assigned to a mass storage device
pattern - the general sequence in which the data structure is accessed (eg.
from tuple to tuple, from record to record, from segment to segment, etc.)
plan - the control structure produced during program preparation and used
by a data base manager to process SQL statements during application execution.
time - the time interval between the instant an instruction initiates a
request for data and the instant the first of the data satisfying the request
is delivered. Note that there is a difference - sometimes large - between the
time data is first delivered and the time when ALL the data is delivered.
- a qualitative assessment of freedom from error or a quantitative measure of
the magnitude of error, expressed as a function of relative error.
data dictionary - a data dictionary that is the sole source for an application
program insofar as metadata is concerned.
- (1) the lowest level function on an activity chart (sometimes called the &atomic
level&) (2) a logical description of a function performed by an enterprise
(3) a procedure (automated or not) designed for the fulfillment of an activity
ratio - the fraction of records in a data base which have activity or are
otherwise accessed in a given period of time or in a given batch run.
processing - one time only, casual access and manipulation of data on parameters
never before used
- an identification (eg., number, name, storage location, byte offset, etc.)
for a location where data is stored
addressing
- the means of assigning data to storage locations, and locating the data upon
subsequent retrieval, on the basis of the key of the data
after image
- the snapshot of data placed on a log upon the completion of a transaction.
change - a motivating force large enough to not be denied, usually aging
of systems, changes in technology, radical changes in requirements, etc.
- Advanced Interactive executive - IBM's version of the UNIX operating system.
- a set of statements organized to solve a problem in a finite number of steps
- an alternative label used to refer to a data element
alphabetic
- a representation of data using letters - upper and/or lower case - only
alphanumeric
- a representation of data using numbers and/or letters, and punctuation
- American National Standards Institute
analytical
processing - the usage of the computer to produce an analysis for management
decision, usually involving trend analysis, drill down analysis, demographic
analysis, profiling, etc.
anticipatory
staging - the technique of moving blocks of data from one storage device
to another with a shorter access time, in anticipation of their being needed
by a program in execution or a program soon to go into execution.
- Application Program Interface - the common set of parameters needed to connect
the communications between programs.
application
- a group of algorithms and data interlinked to support an organizational requirement
application
blocking of data - the grouping into the same physical unit of storage multiple
occurrences of data controlled at the application level
application
data base - a collection of data organized to support a specific application
data base - a collection of data containing data of a historical nature.
As a rule, archival data cannot be updated. Each unit of archival data is relevant
to a moment in time, now passed.
- in network data bases, a named collection of records that can contain occurrences
of one or more record types. A record type can occur in more than one area.
- a design technique used to represent referential integrity in the DSS environment.
artificial
intelligence - the capability of a system to perform functions typically
associated with human intelligence and reasoning.
association
- a relationship between two entities that is represented in a data model
associative
storage - (1) a storage device whose records are identified by a specific
part of their contents rather than their name or physical position in the data
base (2) Content addressable memory. See also parallel search storage.
- (1) data stored in a data warehouse (2) the lowest level of process analysis
data base - a data base made up of p an enterprise
a DSS foundation data base.
- the property in which a group of actions is invisible to other actions executing
concurrently to yield the effect of serial execution. It is recoverable with
successful completion (i.e., commit) or total backout (i.e., rollback) of previous
changes associated with that group.
level data - data with the lowest level of granularity. Atomic level data
sits in a data warehouse and is time variant (i.e., accurate as of some moment
in time now passed)
- a property that can assume values for entities or relationships. Entities
can be assigned several attributes (eg., a tuple in a relationship consists
of values). Some systems also allow relationships to have attributes as well.
audit trail
- data that is available to trace activity, usually update activity.
authorization
identifier - a character string that designates a set of privilege descriptors.
availability
- a measure of the reliability of a system, indicating the fraction of time
when the system is up and available divided by the amount of time the system
should be up and available. Note there is a difference between a piece of hardware
being available and the systems running on the hardware also being available.
end processor - a data base machine or an intelligent disk controller
- to restore the data base to its state as of some previous moment in time
- a file serving as a basis for the activity of backing up a data base. Usually
a snapshot of a database as of some previous moment in time.
Backus-Naur
Form (BNF) - a metalanguage used to specify or describe the syntax of a
language. In BNF, each symbol on the left side of the forms can be replaced
by the symbol strings on the right to develop sentences in the grammar of the
defined language. Synonymous with Backus - Normal Form.
recovery - a recovery technique that restores a data base to an earlier
state by applying before images.
base relation
- a relation that is not derivable from other relations in the data base
- computer environment in which programs (usually long running, sequentially
oriented) access data exclusively, and user interaction is not allowed while
the activity is occurring.
batch environment
- a sequentially dominate in batch, input is collected
and stored for future, later processing. Once collected, the batch input is
transacted sequentially against one or more data bases.
batch window
- the time at which the online system is available for batch or sequential processing.
The batch window occurs during nonpeak processing hours.
image - a snapshot of a record prior to update, usually placed on an activity
materials - a listing of the parts used in a manufacturing process along
with the relation of one product to another insofar as assembly of the final
product is concerned. The bill of materials is a classical recursive structure.
element - a constituent element of data that takes either of two values
or states - either true or false, or one or zero.
search - a dichotomizing search with steps in which the sets of remaining
items are partitioned into two equal parts.
- (1) to assign a value to a data element, variable, or parameter. (2) the attachment
of a data definition to a program prior to the execution of the program.
time - the moment in time when the data description known to the dictionary
is assigned to or bound to the procedural code.
- (b)inary digi(t) - the lowest level of storage. A bit can be in a 1 state
or a 0 state.
- a specialized form of an index indicating the existence or nonexistence of
a condition for a group of blocks or records. Bit maps are expensive to build
and maintain, but provide very fast comparison and access facilities.
- (1) a basic unit of structuring storage (2) the physical unit of transport
and storage. A block usually contains one or more records (or contains the space
for one or more records). In some DBMS a block is called a page.
- the combining of two or more physical records so that they are physically
colocated together. The result of their physical colocation is that they can
be accessed and fetched by a single execution of a machine instruction.
block splitting
- the data management activity in which a filled block is written into two unfilled
blocks, leaving space for future insertions and updates in the two partially
filled blocks.
- a binary storage structure and access method that maintains order in a data
base by continually dividing possible choices into two equal parts and reestablishing
pointers to the respective sets but not allowing more than two levels of difference
to exist concurrently.
- an area of storage that holds data temporarily in main memory while data is
being transmitted, received, read, or written. A buffer is often used to compensate
for the differences in the timing of transmission and execution of devices.
Buffers are used in terminals, peripheral devices, storage units, and CPUs.
- the hardware connection that allows data to flow from one component to another
(eg., from the CPU to the line printer.)
- a basic unit of storage - made up of 8 bits.
- a programming language.
- a buffer usually built and maintained at the device level. Retrieving data
out of a cache is much quicker than retrieving data out of a cylinder.
- to invoke the execution of a module.
model - a data model that represents the inherent structure of data without
regard to either individual use or hardware or software implementation.
cardinality
(of a relation) - the number of tuples (i.e., rows) in a relation. See also
degree of a relation.
- Computer Aided Software Engineering
- a directory of all files available to the computer.
- an organization in which records or other items of data are strung together.
chain list
- a list in which the items cannot be located in sequence but in which each
item contains an identifier (i.e., pointer) for the finding of the next item.
- a subsystem for input and output to and from the computer. Data from storage
units, for example, flows into the computer by way of a channel.
- a member of the standard set of elements used to represent data in the data
type - the characters that can represent the value of an attribute.
checkpoint
- an identified snapshot of the database or a point at which the transactions
against the data base have been frozen or have been quiesced.
checkpoint/restart
- a means of restarting a program at some point other than the beginning - for
example, when a failure or interruption has occurred. N checkpoints may be used
at intervals throughout an application program. At each of those points sufficient
information is stored to permit the program to be restored to the moment in
time the checkpoint has been taken.
- a unit of data existing in a 1:n relationship with another unit of data called
a parent where the parent must exist before the child can exist, but the parent
can exist even if no child unit of data exists.
- chief information officer - an organizational position managing all of the
information processing functions.
file (queue) - an organization of data in which a finite number of units
of data are allocated. Data is then loaded into those units. Upon reaching the
end of the allocated units, new data is written over older data at the start
of the queue. Sometimes called a &wrap around& queue.
- Customer Information Control System - an IBM teleprocessing monitor.
- the facetiously named system development life cycle for analytical, DSS systems.
CLDS is so named because in fact it is the reverse of the classical systems
development life cycle - SDLC.
block - a second or subsequent physical block of data designated to store
table data when the originally allocated block has run out of space.
entities) - all possible entities held by a given proposition.
- (1) in Teradata, a group of physical devices controlled by the same AMP (2)
in DB2 and Oracle, the practice of physically colocating data in the same block
based on the content of data.
key - the key around which data is clustered in a block (DB2/Oracle).
- to combine two or more sets of items into any single set.
- Common Business Oriented Language - a computer language for the business world.
A very common language.
model - a network data base model that was originally defined by the Data
Base Task Group (DBTG) of the Conference on Data System Language (CODASYL) organization.
- (1) to represent data or a computer program in a form that can be accepted
by a data processor (2) to transform data so that it cannot be understood by
anyone who does not have the algorithm used to decode the data prior to presentation
(sometimes called &encode&).
- the event that occurs when two or more records of data are assigned the same
physical location. Collisions are associated with randomizers or hashers.
- a vertical table in which values are selected from the same domain. A row
is made up of one or more columns.
- (1) the specification of an activity by the programmer (2) the actual execution
of the specification.
- a condition raised by the programmer signalling to the DBMS that all update
activity done by the program be executed against a data base. Prior to the commit,
all update activity can be rolled back or cancelled with no ill effects on the
contents of the data base.
protocol - an algorithm to ensure that a transaction is successfully completed.
commonality
of data - similar or identical data that occurs in different applications
or systems. The recognition and management of commonality of data is one of
the foundations of conceptual and physical data base design.
communication
network - the collection of transmission facilities, network processors,
and so on, which provides for data movement among terminals and information
processors.
compaction
- a technique for reducing the number of bits required to represent data without
losing the content of the data. With compaction, repetitive data is represented
very concisely.
- a data item or array of data items whose component type defines a collection
of occurrences with the same data type.
index - an index over multiple columns.
concatenate
- to link or connect two strings of characters, generally for the purpose of
using them as a single value.
conceptual
schema - a consistent collection of data structures expressing the data
needs of the organization. This schema is a comprehensive, base level, and logical
description of the environment in which an organization exists, free of physical
structure and application system considerations.
concurrent
operations - activities executed simultaneously, or during the same time
condensation
- the process of reducing the volume of data managed without reducing the logical
consistency of the data. Condensation is essentially different than compaction.
- to forge a relationship between two entities, particularly in a network system.
- a symbol used to indicate that one occurrence of data has a relationship with
another occurrence of data. Connectors are used in conceptual data base design
and can be implemented hierarchically, relationally, in an inverted fashion,
or by a network.
addressable memory - main storage that can be addressed by the contents
of the data in the memory, as opposed to conventional location addressable memory.
contention
- the condition that occurs when two or more programs try to access the same
data at the same time.
continuous
time span data - data organized so that a continuous definition of data
over a span of time is represented by one or more records.
character - a character whose occurrence in a particular context initiates,
modifies, or stops an operation.
data base - a utilitarian data base containing data not directly related
to the application being built. Typical control data bases are audit data bases,
terminal data bases, security data bases, etc.
cooperative
processing - the ability to distribute resources (programs, files and data
bases) across the network.
coordinator
- the two phase commit protocol defines one data base management system as coordinator
for the commit process. The coordinator is responsible to communicate with the
other data base manager involved in a unit of work.
information warehouse (cif) - the architectural framework that houses the
ODS, data warehouse, data marts, i/t interface, and the operational environment.
The cif is held together logically by metadata and physically by a network such
as the Internet.
- central processing unit.
- the state of processing in which the computer can produce no more output because
the CPU portion of the processor is being used at 100% capacity. When the computer
is CPU-bound, typically the memory and storage processing units are less than
100% utilized. With modern DBMS, it is much more likely that the computer be
I/O-bound, rather than CPU-bound.
- Cross System Product - an IBM application generator.
Common User Access. Specifies the ways in which the user interface to systems
is to be constructed.
value data - data whose accuracy is valid as of the moment of execution.
As opposed to time variant data.
- (1) an indicator that designates a current position on a screen (2) a system
facility that allows the programmer to thumb from one record to the next when
the system has retrieved a set of records.
stability - an option that allows data to move under the cursor. Once the
program is through using the data examined by the cursor it is released. As
opposed to repeatable read.
- the area of storage of DASD that can be read without the movement of the arm.
The term originated with disk files, in which a cylinder consisted of one track
on each disk surface so that each of these tracks could have a read/write head
positioned over it simultaneously.
- see direct access storage device.
- a recording of facts, concepts, or instructions on a storage medium for communication,
retrieval, and processing by automatic means and presentation as information
that is understandable by human beings.
data administrator
- (DA) - the individual or organization responsible for the specification,
acquisition, and maintenance of data management software and the design, validation
and security of files or data bases. The data model and the data dictionary
are classically the charge of the DA.
data aggregate
- a collection of data items.
- a collection of interrelated data stored (often with controlled, limited redundancy)
according to a schema. A data base can serve a single or multiple applications.
administrator (DBA) - the organizational function charged with the day to
day monitoring and care of the data bases. The dba function is more closely
associated with physical data base design than the DA is.
key - a unique value that exists for each record in a data base. The value
is often indexed, although it can be randomized or hashed.
machine - a dedicated-purpose computer that provides data access and management
through total control of the access method, physical storage, and data organization.
Often called a &back end processor.& Data is usually managed in parallel
by a data base machine.
management system (DBMS) - a computer based software system used to establish
and manage data.
record - a physical root and all of its dependents (in IMS).
data definition
- the specification of the data entities, their attributes, and their relationships
in a coherent data base structure to create a schema.
data definition
language (DDL) (also called a data description language) - the language
used to define the data base schema and additional data features that allows
the DBMS to generate and manage the internal tables, indexes, buffers, and storage
necessary for data base processing.
data description
language - see data definition language.
data dictionary
- a software tool for recording the definition of data, the relationship of
one category of data to another, the attributes and keys of groups of data,
and so forth.
data division
(COBOL) - the section of a COBOL program that consists of entries used to
define the nature and characteristics of the data to be processed by the program.
data driven
development - the approach to development that centers around identifying
the commonality of data through a data model and building programs that have
a broader scope than the immediate application. Data driven development differs
from classical application oriented development.
data driven
process - a process whose resource consumption depends on the data on which
it operates. For example, a hierarchical root has a dependent. For one occurrence
the are two dependents for the root. For another occurrence of the root there
are 1,000 occurrences of the dependent. The same program that accesses the root
and all its dependents will use very different amounts of resources when operating
against the two roots although the code will be exactly the same.
data element
- (1) an attribute of an entity (2) a uniquely named and well defined category
of data that consists of data items and that is included in a record of an activity.
data engineering
(see information engineering) - the planning and building of data structures
according to accepted mathematical models, on the basis of the inherent characteristics
of the data itself, and independent of hardware and software systems.
data independence
- the property of being able to modify the overall logical and physical structure
of data without changing any of the application code supporting the data.
- a discrete representation having the properties that define the data element
to which it belongs. See data element.
set (dis) - a grouping of data items, each of which directly relates to
the key of the grouping of data in which the data items reside. The data item
set is found in the mid level model.
data manipulation
language (DML) - (1) a programming language that is supported by a DBMS
and used to access a data base (2) language constructs added to a higher-order
language (eg., COBOL) for the purpose of data base manipulation.
- a department specific data warehouse. There are two types of data marts
- independent and dependent. An independent data mart is fed data directly from
the legacy environment. A dependent data mart is fed data from the enterprise
data warehouse. In the long run, dependent data marts are architecturally much
more stable than independent data marts.
data model
- (1) the logical data structures, including operations and constraints provided
by a DBMS for effective data base processing (2) the system used for the representation
of data (eg., the ERD or relational model.)
data record
- an identifiable set of data values treated as a unit, an occurrence of a schema
in a data base, or a collection of atomic data items describing a specific object,
event, or tuple.
data security
- the protection of the data in a data base against unauthorized disclosure,
alteration, or destruction. There are different levels of security.
- a named collection of logically related data items, arranged in a prescribed
manner, and described by control information to which the programming systems
has access.
data storage
description language (DSDL) - a language to define the organization of stored
data in terms of an operating system and device independent storage environment.
See also device media control language.
data structure
- a logical relationship among data elements that is designed to support specific
data manipulation functions (eg., trees, lists, and tables.)
- the definition of a set of representable values that is primitive and without
meaningful logical subdivision.
- see user view.
data volatility
- the rate of change of the content of data.
data warehouse
- a collection of integrated subject oriented data bases designed to support
the DSS function, where each unit of data is relevant to some moment in time.
The data warehouse contains atomic data and lightly summarized data. A data
warehouse is a subject oriented, integrated, non volatile, time variant collection
of data designed to support management DSS needs
data warehouse
administrator (dwa) - the organization function designed to create and maintain
the data warehouse. The dwa combines several disciplines, such as the da, dba,
emn user, etc.
- a data base management system by IBM.
- data base / data communications
DBMS language
interface (DB I/O module) - software that applications invoke in order to
access a data base. The module in turn has direct access with the DBMS. Standards
enforcement and standard error checking are often features of an I/O module.
- see deadly embrace.
embrace - the event that occurs when transaction A desires to access data
currently protected by transaction B, while at the same time transaction B desires
to access data that is currently being protected by transaction A. The deadly
embrace condition is a serious impediment to performance.
support system (DSS) - a system used to support managerial decisions. Usually
DSS involves the analysis of many units of data in a heuristic fashion. As a
rule, DSS processing does not involve the update of data.
decompaction
- the op once data is stored in a compacted form, it must
be decompacted to be used.
decryption
- the opposite of encryption. Once data is stored in an encrypted fashion, it
must be decrypted in order that it can be used.
(of a relation) - the number of attributes or columns of a relation. See
cardinality of a relation.
- a flag, symbol, or convention used to mark the boundaries of a record, field,
or other unit of storage.
staging - the movement of blocks of data from one storage device to another
device with a shorter access time when programs request the blocks and the blocks
are not already in the faster access storage.
denormalization
- the technique of placing normalized data in a physical location that optimizes
the performance of the system.
data - data whose existence depends on two or more occurrences of a major
subject of the enterprise.
data element - a data element that is not necessarily stored but that can
be generated when needed (eg., age given current date and date of birth.)
relation - a relation that can be obtained from previously defined relations
by applying some sequence of retrieval and derivation operator (eg., a table
that is the join of others plus some projections.) See a virtual relation.
review - the quality assurance process in which all aspects of a system
are reviewed publicly prior to the striking of code.
media control language (DMCL) - a language used to define the mapping of
the data onto the physical storage media. See data storage description language.
table - the table that is joined to a fact table in a star join. The dimension
table is the structure that represents then non populous occurrences of data
in a data mart.
access - retrieval or storage of data by reference to its location on a
volume. The access mechanism goes directly to the data in question, as is generally
required with online use of data. Also called random access or hashed access.
access storage device (DASD) - a data storage unit on which data can be
accessed directly without having to progress through a serial file such as a
magnetic tape file. A disk unit is a direct access storage device.
- a table specifying the relationships between items of data. Sometimes a table
or index giving the addresses of data.
distributed
catalog - a distributed catalog is needed to achieve site autonomy. The
catalog at each site maintains information about objects in the local data bases.
The distributed catalog keeps information on replicated and distributed tables
stored at that site and information on remote tables located at another site
that cannot be accessed locally.
distributed
data base - a data base controlled by a central DBMS but in which the storage
devices are geographically dispersed or not attached to the same processor.
See parallel I/O.
distributed
data warehouse - where more than one enterprise data warehouse is built,
the combination is called a distributed data warehouse
distributed
environment - a set of related data processing systems, where each system
has its own capacity to operate autonomously, but with some applications which
execute at multiple sites. Some of the systems may be connected with teleprocessing
links into a network in which each system is a node.
distributed
free space - space left empty at intervals in a data layout to permit insertion
of new data.
distributed
metadata - distributed metadata is data metadata that resides at different
architectural entities, such as data marts, enterprise data warehouses, ODS,
distributed
request - a transaction across multiple nodes
distributed
unit of work - the work done by a transaction that operates against multiple
- an operation that partitions a relation on the basis of the contents of data
found in the relation.
- IBM's Data Language One, for describing logical and physical data structures
- the set of legal values from which actual values are derived for an attribute
or a data element.
data - data loaded into a data warehouse that has a future probability of
access of zero
- the stripping of data from one data base to another based on the content of
data found in the first data base.
drill down
analysis - the type of analysis where examination of a summary number leads
to the exploration of the components of the sum.
base - the practice of separating high performance, transaction oriented
data from decision support data
base management systems - the practice of using multiple data base management
systems to control different aspects of the data base environment
dumb terminal
- a device used to interact directly with the end user where all processing
is done on a remote computer. A dumb terminal acts as a device that gathers
data and displays data only.
SQL - SQL statements that are prepared and executed within a program while
the program is executing. In dynamic SQL the SQL source is contained in host
language variables rather than being coded into the application program.
storage allocation - a technique in which the storage areas assigned to
computer programs are determined during processing.
subset of data - a subset of data selected by a program and operated on
only by the program, and released by the program once the program ceases execution.
- Electronic Data Interchange.
EIS (Executive
Information Systems) - systems designed for the top executive, featuring
drill down analysis and trend analysis.
pointer - a record pointer (i.e., a means of internally linking related
records) that is not available to an external index or directory. Embedded pointers
are used to reduce search time, but require maintenance overhead.
- a shortening or abbreviation of the physical representation of a data value
(eg., male = &M&, female =&F&)
encryption
- the transformation of data from a recognizable form to a form unrecognizable
without the algorithm used for the encryption. Encryption is usually done for
the purposes of security.
enterprise
- the generic term for the company, corporation, agency, or business unit. Usually
associated with data modelling.
enterprise
data warehouse - a data warehouse holding the most atomic data the corporation
has. Two or more enterprise data warehouses may be combined in order to create
a distributed data warehouse
- a person, place or thing of interest to the data modeller at the highest level
of abstraction.
- relationship - attribute (ERA) model - a data model that defines entities,
the relationship between the entities, and the attributes that have values to
describe the properties of entities and/or relationships.
- relationship diagram (ERD) - a high level data model - the schematic showing
all the entities within the scope of integration and the direct relationship
between those entities.
- a signal that an activity of significance has occurred. An event is noted
by the information system.
event discrete
data - data relating to the measurement or description of an event.
system - a system that captures and automates the usage of human experience
and intelligence.
- a DSS end user who operates on a random basis looking at large amounts of
detailed data for patterns, associations, and other previously unnoticed relationships
- (1) a list of unsigned integers that specifies an array (2) a physical unit
of disk storage attached to a data set after the initial allocation of data
has been made.
data - (1) data originating from other than the operational systems of a
corporation (2) data residing outside the central processing complex
schema - a logical description of a user's method of organizing and structuring
data. Some attributes or relationships can be omitted from the corresponding
conceptual schema or can be renamed or otherwise transformed. See view.
- the process of selecting data from one environment and transporting it to
another environment.
table - the central component of the star join. The fact table is the structure
where the vast majority of the occurrences of data in the data mart reside
- a DSS user who repetitively looks at small amounts of data and who often finds
what he/she is looking for
See data item.
- a set of related records treated as a unit and stored under a single logical
file name.
first in first out - a fundamental ordering of processing in a queue.
first in last out - a standard order of processing in a stack.
- an indicator or character that signals the occurrence of some condition.
- a collection of records containing no data aggregates, nested repeated data
items, or groups of data items.
disk - a device for storing data on a personal computer.
key - an attribute that is not a primary key in a relational system, but
whose values are the values of the primary key of another relation.
- the arrangement or layout of data in or on a data medium or in a program definition.
recovery - a recovery technique that restores a data base by reapplying
all transactions using a before image from a specified point in time to a copy
of the data base taken at that moment in time.
generation language - language or technology designed to allow the end user
unfettered access to data.
functional
decomposition - the division of operations into hierarchical functions (i.e.,
activities) that form the basis for procedures.
- a measurement of data between a megabyte and a terabyte - 10 x 10**9 bytes
data warehouse - a data warehouse that is distributed around the world.
In a global data warehouse the system of record resides in the local site.
- a symbol produced on a screen representing an object or a process in the real
granularity
- the level of detail contained in a unit of data. The more detail there is,
the lower the level of granularity. The less detail there is, the higher the
level of granularity.
- to convert the value of the key of a record into a location on DASD.
hash total
- a total of the values of one or more fields, used for the purposes of auditability
and control.
record or header table - a record containing common, constant or identifying
information for a group of records that follow.
- the mode of analysis in which the next step is determined by the results of
the current step of analysis. Used for decision support processing.
hierarchical
model - a data model providing a tree structure for relating data elements
or groups of data elements. Each node in the structure represents a group of
data elements or a record type. There can be only one root node at the start
of the hierarchical structure.
- an occurrence of data that satisfies some search criteria.
- a measure of the number of records in a file expected to be accessed in
a given run. Usually expressed as a percentage - number of input transactions/number
of records in the file x 100 = hit ratio
- identical names that refer to different attributes.
horizontal
distribution - the splitting of a table across different sites by rows.
With horizontal distribution rows of a single table reside at different sites
in a distributed data base network.
- the processor receiving and processing a transaction.
code - a code for data compaction in which frequently used characters are
encoded with fewer bits than infrequently used characters.
- a network DBMS from CA.
- Institute of Electrical and Electronics Engineers.
IMS - Information
Management System - an operational DBMS by IBM.
image copy
- a procedure in which a data base is physically copied to another medium
for the purposes of backup.
- the portion of the storage structure maintained to provide efficient access
to a record when its index key item is known.
index chains
- chains of data within an index.
index sequential
access method (ISAM) - a file structure and access method in which records
can be processed sequentially (eg., in order, by key) or by directly looking
up their locations on a table, thus making it unnecessary to process previously
inserted records.
index point
- a hardware reference ma used for timing purposes.
addressing - any method of specifying or locating a record through calculation
(eg., locating a record through the scan of an index)
information
- data that human beings assimilate and evaluate to solve a problem or make
a decision.
information
engineering (IE) - the discipline of creating a data driven development
environment.
information
center - the organizational unit charged with identifying and accessing
information needed in DSS processing.
- a leading data warehouse vendor
input/output
(I/O) - the means by which data is stored and/or retrieved on DASD. I/O
is measured in milliseconds (i.e., mechanical speeds) whereas computer processing
is measured in nanoseconds (i.e., electronic speeds).
- a set of values representing a specific entity belonging to a particular entity
type. A single value is also the instance of a data item.
integration/transformation
(i/t) program - a program designed to convert and move data from the legacy
environment to the data warehouse environment. I/T programs are notoriously
unstable and require constant maintenance
- the property of a data base that ensures that the data contained in the data
base is as accurate and consistent as possible.
intelligent
data base - a data base that contains shared logic as well as shared data
and automatically invokes that logic when the data is accessed. Logic, constraints,
and controls relating to the use of the data are represented in an intelligent
data model.
interactive
- a mode of processing that combines some of the characteristics of online transaction
processing and batch processing. In interactive processing the end user interacts
with data over which he/she has exclusive control. In addition, the end user
can initiate background activity to be run against the data.
interleaved
data - data from different tables mixed into a simple table space where
there is commonality of physical colocation based on a common key value.
schema - the schema that describes logical structures of the data and the
physical media over which physical storage is mapped.
- a network that connects many public users
interpretive
- a mode of data manipulation in which the commands to the DBMS are translated
as the user enters them (as opposed to the programed mode of process manipulation.)
intersection
data - data that is associated with the junction of two or more record types
or entities, but which has no meaning when disassociated with any records or
entities forming the junction.
- a network that connects many private users
file - a file structure that uses an inverted index, where entries are grouped
according to the content of the key being referenced. Inverted files provide
for the fast spontaneous searching of files.
index - an index structure organized by means of a nonunique key to speed
the search for data by content.
list - a list organized around a secondary index instead of around a primary
- input / output operation. Input / output operations are the key to performance
because they operate at mechanical speeds, not at electronic speeds.
- the point after which no more processing can be done because the I/O subsystem
is saturated.
- see Indexed Sequential Access Method.
a type of& - an analytical tool used in abstracting data during the
process of conceptual data base design (eg., a cocker spaniel is a type of dog.)
ISDN (Integrated
Services Digital Network) - telecommunications technology that enables companies
to transfer data and voice through the same phone lines.
- international Standards Organization
see data item.
- a classification of an item according to its domain, generally in a gross
analysis - the mode of processing in which the next step of processing depends
on the results obtained by the existi heuristic processing.
(joint application design) - an organization of people - usually end users
- to create and refine application system requirements.
- an operation that takes two relations as operands and produces a new relation
by concatenating the tuples and matching the corresponding columns when a stated
condition holds between the two.
sample - a sample of data where data is accepted or rejected for the sample
based on one or more parameters.
- from the network environment, an occurrence of data that has two or more parent
segments. For example, an order for supplies must have a supplier parent and
a part parent.
- to adjust the value representation in a character field to the right or to
the left, ignoring blanks encountered.
- a sequence of data base keys maintained by the DBMS for the duration of the
- a data item or combination of data items used to identify or locate a record
instance (or other similar data groupings.)
key compression
- a technique for reducing the nu used in making indexes
occupy less space.
key, primary
- a unique attribute used to identify a single record in a data base.
key, secondary
- a nonunique attribute used to identify a class of records in a data base.
- a set of symbols used to identify or describe an item, record, message, or
file. Occasionally a label may be the same as the address of the record in storage.
- a set of characters, conventions, and rules used to convey information and
consisting of syntax and semantics.
- the time taken by a DASD device to position the read arm over the physical
storage medium. For general purposes, average latency time is used.
least frequently
used (LFU) - a replacement strategy in which new data must replace existing
data i the least frequently used items are replaced.
least recently
used (LRU) - a replacement strategy in which new data must replace existing
data i the least recently used items are replaced.
environment - the transaction oriented, application based environment
abstraction - the level of abstraction appropriate to a dimension. The level
of abstraction which is appropriate is entirely dependent on the ultimate user
of the system.
- the hardware by which data flows to or from the processor. Lines typically
go to terminals, printers, and other processors.
line polling
- the activity of the teleprocessing monitor in which different lines are queried
to determine whether they have data and/or transactions that need to be transmitted.
- the length of time required for a transaction to go from either the terminal
to the processor or the processor to the terminal. Typically line time is the
single largest component of online response time.
- the ability to relate one unit of data to another.
list - set of records in which each record contains a pointer to the next
record on the list. See chain.
- an ordered set of data items.
sample - a representative data base typically used for heuristic statistical
analytical processing in place of a large data base. Periodically the very large
data base is selectively stripped of data so that the resulting living sample
data base represents a cross section of the very large data base as of some
moment in time.
- to insert data values into a data base that was previously empty.
local site
support - within a distributed unit of work, a local site update allows
a process to perform SQL update statements referring to the local site.
local transaction
- in a distributed DBMS, a transaction that requires reference only to data
that is stored at the site where the transaction originated.
of processing - in distributed data base, the design of processing so that
remote access of data is eliminated or reduced substantively.
- the event that occurs when update is done against a data base record and the
transaction has not yet reached a commit point. The online transaction needs
to prevent other transactions from accessing the data while update is occurring.
- a journal of activity.
- the automatic recording of data with regard to the access of the data, the
updates to the data, etc.
representation - a data view or description that does not depend on a physical
storage device or a computer program.
identity - when data is brought in from an external source and the identity
of the external source is discarded, loss of identity occurs. A common practice
with microprocessor data.
Logical Unit Type 6.2 - peer to peer data stream with network operating system
for program to program communication. LU6.2 allows mid-range machines to talk
to one another without the involvement of the mainframe.
learning - the ability of a machine to improve its performance automatically
based on past performance.
tape - (1) the storage medium most closely associated with sequential processing
(2) a large ribbon on which magnetic images are stored and retrieved.
main storage
data base (msdb) - a data base that resides entirely in main storage. Such
data bases are very fast to access, but require special handling at the time
of update. Another limitation of msdb's are that they can only manage small
amounts of data.
file - a file that holds the system of record for a given set of data (usually
bound by an application.)
transaction arrival rate (MTAR) - the rate of arrival of transactions at
the moment of peak period processing.
- a measurement of data - 10 x 10**6 bytes of data
- (1) the data input by the user in the online environment that is used to drive
a transaction (2) the output of a transaction.
- (1) data about data (2) the description of the structure, content, keys, indexes,
etc. of data
metalanguage
- a language used to specify other languages.
microprocessor
- a small processor serving the needs of a single user.
- the process by which frequently used items of data are moved to more readily
accessible areas of storage and infrequently used items of data are moved to
less readily accessible areas of storage.
mips (million
instructions per second) - the standard measurement of processor speed for
minicomputers and mainframe computers.
operation - a classification for systems that execute in a similar fashion
and share distinctive operational characteristics. Some modes of operation are
operational, DSS, online, interactive, etc.
- an arithmetic term describing the remainder of a division process. 10 modulo
7 is 3. Modulo is usually associated with the randomization process.
organization - a chained file organization in which the chains are divided
into fragments and each fragment is indexed. This organization of data permits
faster access to the data.
key retrieval - that requires searches of data on the basis of the values
of several key fields (some or all of which are secondary keys.)
MVS - Multiple
Virtual Storage - IBM's mainline operating system for mainframe processors.
There are several extensions of MVS.
Pipes - program to program protocol with Microsoft's LAN manager. The Named
Pipes API supports intra and inter machine process to process communications.
first normal
form - data that has been organized into two dimensional flat files without
repeating groups
normal form - data that functionally depends on the entire candidate key
third normal
form - data that has had all transitive dependencies on data items other
than the candidate key removed.
normal form - data whose candidate key is related to all data items in the
record and that contains no more than one nontrivial multivalued dependency
on the candidate key.
join - a join in which the redundant logic components generated by the join
are removed.
language - a language generally spoken, whose rules are based on current
usage and not explicitly defined by a grammar.
- to steer a course through a data base, from record to record, by means of
an algorithm which examines the content of data.
- a computer network consists of a collection of circuits, data switching elements
and computing systems. The switching devices in the network are called communication
processors. A network provides a configuration for computer systems and communication
facilities within which data can be stored and accessed and within which DBMS
can operate.
model - a data model that provides data relationships on the basis of records,
and groups of records (i.e., sets) in which one record is designated as the
set owner, and a single member record can belong to one or more sets.
complement - transformation of a numeric field calculated by subtracting
the initial value from a filed consisting of all nines.
- a point in the network at which data is switched.
nonprocedural
language - syntax that directs the computer as to what to do, not how to
do it. Typical non procedural languages include RAMIS, FOCUS, NOMAD, and SQL.
- to decompose complex data structures into natural structures.
- an operating system built by Microsoft
- an item or record for which no value currently exists or possibly may ever
- a representation using only numbers and the decimal point.
occurrence
- see instance.
pointer - an indirect pointer. An offset pointer exists inside a block and
the index points to the offset. If data must be moved, only the offset pointer
in the b the index entry remains untouched.
storage - storage devices and storage medium where data can be accessed
in a direct fashion.
operational
data store (ODS) - the form that data warehouse takes in the operational
environment. Operational data stores can be updated, do provide rapid and consistent
response time, and contain only a limited amount of historical data.
operational
application - see legacy application
system - software that enables a computer to supervise its own operations
and automatically call in programs, routines, languages, and data as needed
for continuous operation throughout the execution of different types of jobs.
operational
data - data used to support the daily processing a company does.
operations
- the department charged with the running of the computer.
disk - a storage medium using lasers as opposed to magnetic devices. Optical
disk is typically write only, is much less expensive per byte than magnetic
storage, and is highly reliable.
- a DBMS by ORACLE Corp.
- to place items in an arrangement specified by such rules as numeric or alphabetic
order. See sort.
- the operating system for IBM's Personal System / 2.
- Open Software Foundation
- Open Systems Interconnection
- (1) the condition in which a record or a segment cannot be stored in its home
address because the address is already occupied. In this case the data is placed
in another location referred to as overflow. (2) the area of DASD where data
is sent when the overflow condition is triggered.
- the responsibility for update for operational data.
- a technique used to fill a field, record, or block with default data (eg.,
blanks or zeros)
- (1) a basic unit of data on DASD (2) a basic unit of storage in main memory.
page fault
- a program interruption that occurs when a page that is referred to is not
in main memory and must be read in from external storage.
page fixed
- the state in which programs or data cannot be removed from main storage.
Only a limited amount of storage can be page fixed.
- in virtual storage systems, the technique of making memory appear to be larger
than it really is by transferring blocks (pages) of data or programs into external
data organization - an arrangement of data in which the data is spread over
independent storage devices and is managed independently.
I/O - the process of accessing or storing data on multiple physical data
search storage - a storage device in which one or more parts of all storage
locations are queried simultaneously for a certain condition or under certain
parameters. See associative storage.
- an elementary data value used as a criteria for qualification, usually of
searches of data or in the control of modules.
- a unit of data in a 1:n relationship with another unit of data called a child,
where the parent can exist independently, but the child cannot exist unless
there is a parent.
- the algorithm that translates syntax into meaningful machine instructions.
Parsing determines the meaning of statements issued in the data manipulation
- a segmentation technique in which data is divided into physically different
units. Partitioning can be done at the application or the system level.
path length
- the number of instructions executed for a given program or instruction.
peak period
- the time when the most transactions arrive at the computer with the expectation
of execution.
performance
- the length of time from the moment a request is issued until the first of
the results of the request are received.
discrete data - a measurement or description of data taken at a regular
time interval.
representation - (1) the representation and storage of data on a medium
such as magnetic storage (2) the description of data that depends on such physical
factors as length of elements, records, pointers, etc.
- vehicles for passing data from one application to another.
network structure - a relationship between records or other groupings of
data in which a child record can have more than one parent record.
plug compatible
manufacturer (PCM) - a manufacturer of equipment that functionally is identical
to that of another manufacturer (usually IBM).
- the address of a record or other groupings of data contained in another record
so that a program may access the former record when it has retrieved the latter
record. The address can be absolute, relative, or symbolic, and hence the pointer
is referred to as absolute, relative, or symbolic.
- the buffers made available to the online controller.
- to place occurrences of data values in a previously empty data base. See load.
- the degree of discrimination with which a quantity is stated. For example
a three digit numeral discriminates among 1,000 possibilities, from 000 to 999.
precompilation
- the processing of source text prior to compilation. In an SQL environment,
SQL statements are replaced with statements that will be recognized by the host
language compiler.
data - data in a segment or a record used exclusively for system control,
usually unavailable to the user.
key - an attribute that contains values that uniquely identify the record
in which the key exists.
data - data whose existence depends on only a single occurrence of a major
subject area of the enterprise.
- the prevention of unauthorized access and manipulation of data.
descriptor - a persistent object used by a DBMS to enforce constraints on
operations.
data base - the component of a DSS application where previously defined
decision parameters are stored. A problems data base is consulted to review
characteristics of past decisions and to determine ways to meet current decision
making needs.
- the hardware at the center of execution of computer programs. Generally speaking
processors are divided into three categories - mainframes, minicomputers, and
microcomputers.
cycles - the hardware's internal cycles that drive the computer (eg., initiate
I/O, perform logic, move data, perform arithmetic functions, etc.)
production
environment - the environment where operational, high performance processing
area - the portion of main memory in which application programs are executed.
progressive
overflow - a method of handling overflow in a randomly organized file that
does not require the use of pointers. An overflow record is stored in the first
available space and is retrieved by a forward serial search from the home address.
projection
- an operation that takes one relation as an operand and returns a second relation
that consists of only the selected attributes or columns, with duplicate rows
eliminated.
proposition
- a statement about entities that asserts or denies that some condition holds
for those entities.
- the call format used by a teleprocessing monitor.
cards - an early storage medium on which data and input were stored. Today
punched cards are rare.
purge data
- the data on or after which a storage area may be overwritten. Used in conjunction
with a file label, it is a means of protecting file data until an agreed upon
release date is reached.
language - a language that enables an end user to interact directly with
a DBMS to retrieve and possibly modify data managed under the DBMS
- an aggregation of values of data organized by their relation to a common key
record-at-a-time
processing - the access of data a record at a time, a tuple at a time, etc.
- the restoration of the database to an original position or condition, often
after major damage to the physical medium
- a DSS data base management system by Red Brick corp
redundancy
- the practice of storing more than one occurrence of data. In the case where
data can be updated, redundancy poses serious problems. In the case where data
is not updated, redundancy is often a valuable and necessary design tool.
referential
integrity - the facility of a DBMS to ensure the validity of a predefined
relationship.
reorganization
- the process of unloading data in a poorly organized state and reloading the
data in a well organized state. Reorganization in some DBMS is used to restructure
data. Reorganization is often called - &reorg&, or an &unload/reload&
groups - a collection of data that can occur several times within a given
record occurrence
summary - a form of storing archival data where the most recent data has
the lowest level of detail stored and the older data has higher levels of detail
of integration - the formal definition of the boundaries of the system being
- system development life cycle - the classical operational system development
life cycle that typically includes requirements gathering, analysis, design,
programming, testing, integration, and implementation. Sometimes called a &waterfall&
development life cycle.
sequential
file - a file in which records are ordered according to the values of one
or more key fields. The records can be processed in this sequence starting from
the first record in the file, continuing to the last record in the file.
file - a sequential file in which the records are physically adjacent, in
sequential order.
set-at-a-time
processing - access of data by groups, each member of which satisfies some
selection criteria
- a database dump or the archiving of data as of some one moment in time
snow flake
structure - the grouping together of two or more star joins
- a denormalized form of organizing data optimal for the access of a group of
people, usually a department. Star joins are usually associated with data marts.
Star joins were popularized by Ralph Kimball
hierarchy - storage units linked to form a storage subsystem, in which some
units are fast to access and consume small amounts of storage, but which are
expensive, and other units are slow to access and are large, but are inexpensive
database - a database organized around a major subject of the corporation.
Classical subject databases are for customer, transaction, product, part, vendor,
- a data base management system by Sybase Corp
log - an audit trail of relevant system events (for example, transaction
entries, database changes, etc.)
of record - the definitive and singular source of operational data or metadata.
If data element abc has a value of 25 in a database record but a value of 45
in the system of record, by definition the first value must be incorrect. The
system of record is useful for the management of redundancy of data. For metadata,
at any one moment in time, each unit of metadata is owned by one and only one
organizational unit.
- a relation that consists of a set of columns with a heading and a set of rows
(i.e., tuples)
- a data base management system by NCR
- a measurement of a large amount of data, 10 x 10**12 bytes
time stamping
- the practice of tagging each record with some moment in time, usually when
the record was created or when the record was passed from one environment to
time variant
data - data whose accuracy is relevant to some one moment in time. The three
common forms of time variant data are continuous time span data, event discrete
data, and periodic discrete data. See current value data.
transaction
processing - the activity of executing many short, fast running programs,
providing the end user with consistent two to three second response time
transition
data - data possessing both primitive and der usually
very sensitive to the running of the business. Typical transition data include
interest rates for a bank, policy rates for an insurance company, retail sale
prices for a manufacturer/distributor, etc.
trend analysis
- the process of looking at homogeneous data over a spectrum of time
true archival
data - data at the lowest level of granularity in the current level detail
- a popular operating system for data warehouses
- to change, add, delete, or replace values in all or selected entries, groups,
or attributes stored in a database
- a person or process issuing commands or messages and receiving stimuli from
the information system
framework - a specification for organizing blueprints for information systems,
popularized the great John Zachman
Copyright (C) by W. H. Inmon, all
rights reserved.