ISO/IEC 14756:1999 情報技術  —  コンピュータベースのソフトウェアシステムのパフォーマンスの測定と評価 | ページ 6

4.1

activity

An order submitted to the system under test (SUT) by a user or an emulated user demanding the execution of a data processing operation according to a defined algorithm to produce specific output data from specific input data and (if requested) stored data.

4.2

activity type

A classification of activities defined by the execution of the same algorithm.

4.3

chain

One or more tasks submitted to the SUT in a defined sequence.

4.4

chain type

A classification of chains which is defined by the sequence of tasks types.

4.5

computer-based software system

CBSS

A software system running on a computer.

Note 1 to entry: A CBSS may be a data processing system as seen by human users at their terminals or at equivalent machine-user-interfaces. It includes hardware and all software (system software and application software) which is necessary for realizing data processing functions required by its users.

4.6

emulated user

The imitation of a user, with regard to the tasks he submits and his time behaviour, realized by a technical system.

4.7

execution time

The time which elapses between task submission and completion.

4.8

mean execution time

The mean value of all execution times of tasks of the j-th task type which were submitted within the rating interval.

Note 1 to entry: The mathematical symbol is T_ME (j) corresponding to the j-th task type.

4.9

mean execution time rating value

The quotient (corresponding to the j-th task type) of the mean execution time reference value and the measured mean execution time.

Note 1 to entry: The mathematical symbol is R_ME (j) corresponding to the j-th task type.

4.10

mean execution time reference value

The mean execution time maximally accepted by the emulated user. It shall be computed from the workload parameter set (see 9.1.1, 15.2.1 and clause B.1).

4.11

observation period

The time interval ここで, the measurement procedure is observed for collecting (logging) measurement results for rating or validation, consisting of the rating interval and the supplementary run.

4.12

preparation time

The time which elapses before the task submission. The event of starting the preparation time depends on the definition of the task mode of the following task. The preparation time value is the random chosen representation of an distributed variable with a defined mean and a standard deviation. They depend on the task type of the following task and the type of the emulated user generating the task.

4.13

rating interval

The time interval of the measurement procedure from the time the SUT reaches a stable state of operation to the time the measurement results are fulfilling the required statistical significance.

4.14

relative chain frequency

The relative frequency of using the I-th chain type by an emulated user of the i-th type.

4.15

remote terminal emulator

RTE

A data processing system realizing a set of emulated users.

4.16

stabilization phase

The time interval of the measurement procedure when the RTE starts submitting tasks until the SUT reaches a stable state of operation.

4.17

supplementary run

The time interval of the measurement procedure from the time the measurement results fulfil the required statistical significance to the time when all tasks, which were submitted during the rating interval, are completed.

4.18

system under test

SUT

The parts of the CBSS to be tested. All components which may influence the SUT's time behaviour shall be part of the SUT and if the influence depends on some workload, this workload shall be represented by the RTE too. Except if an influence to the time behaviour is impossible or not evident the emulation of the parts beside may be omitted.

4.19

task

The combination of:

4.20

task completion

Timely event when for a specific task the total output string or, in case of a set of output strings, all parts are completely received by to the emulated user or another instance. If the task does not submit an output to the user (i.e.: during the measurement: to the RTE) a functionality, producing an „artificial output", indicating the task completion, may be added to the task for usage during the measurement.

4.21

task mode

Indication of whether the user's preparation time begins immediately with the task submission of the preceding task (value = 0, i.e."NO WAIT") or begins when the preceding task has been completed (task completion) (value = 1, i.e."WAIT").

4.22

task submission

Timely event when the input string is completely submitted from the emulated user to the SUT and the execution of the task may start, regardless if the SUT starts the execution immediately or not. The task submission is also indicated by the time when the action of the emulated user for this tasks ends. Following this International Standard, the task submission shall not be defined (and measured) by the event, when the SUT has received or interpreted the input string or when a receipt string, which may be send by the SUT after receiving and interpreting the input string, is received by the emulated user.

Note 1 to entry: Normally the task submission is defined internally by the submission of a special character (e.g. Carriage Return) or a character sequence at the end of the input string or at the end of several parts of the input string. Also it often happens that the task submission event is defined by the submission of the last character of an specified number characters in a string. For a classic batch task the task submission may be defined by the submission of the last character of the last string of the batch command sequence.

4.23

task type

Note 1 to entry: Emulated users submit only these types of tasks to the SUT.

Note 2 to entry: The mathematical symbol for the current number of the task type is j. The total number of task types is m.

4.24

throughput

The rate (i.e. the average number per time unit with respect to the rating interval) of all tasks of a task type submitted to the SUT.

Note 1 to entry: The mathematical symbol is B(j) corresponding to the j-th task type.

Note 2 to entry: Usually throughput is defined by the rate of terminated tasks during a period of time. In order to specify a supplementary run and not a"heads run" in this standard, throughput is defined in this standard by the rate of tasks submitted. Nevertheless the correct task completion is validated by checking the computational result of the task.

4.25

throughput rating value

The quotient (corresponding to the j-th task type) of the (actual) throughput and the throughput reference value.

Note 1 to entry: The mathematical symbol is R_th(j) corresponding to the j-th task type.

4.26

throughput reference value

The minimum throughput required by the set of emulated users.

Note 1 to entry: The mathematical symbol is B_Ref(j) corresponding to the j-th task type.

4.27

time class

A time limit, combined with a relative frequency corresponding to the ratio of the number of tasks (of a specific task type) - with an execution time less than or equal to the corresponding time limit - to the total amount of tasks (of that particular task type), used for comparison with the execution time of a task (of that particular task type).

Note 1 to entry: The mathematical symbol for the total number of time classes of the j-th task type is z(j) ; the time limit is gT(j,c) where j is the current number of the tasktype and c is the current number of the time class, running from 1 to z(j); the relative frequency is r_T(j,c) corresponding to the time limit g_T(j,c).

4.28

timeliness rating value

The quotient (corresponding to the j-th task type) of the timely throughput and the total throughput.

Note 1 to entry: The mathematical symbol is R_ti(j) corresponding to the j-th task type.

4.29

timeliness function

A description of the user requirements with respect to the execution times of tasks of a specific task type. It consists of one or more time classes. The relative frequency of the time class with the highest time limit shall be 1.0 (i.e. = 100%).

Note 1 to entry: The timeliness function corresponding to the j-th task type consists of z(j) time classes, each time class consisting of the pairs of numbers gT(j,c) and r_T(j,c).