1、CEPT T/CAC*Z*E 89 m 2326414 0002906 2 m Recommendation T/CAC 2 (Stockholm 1989 (CAC) T/CAC 2 E Page 1 INDICATORS FOR THE NETWORK PERFORMANCE ASPECTS OF THE QUALITY OF SERVICE OF INTERNATIONAL PACKET SWITCHED SERVICES Recommendation proposed by the project team X.25 Text of the Recorrtmendntion adopt
2、ed by the “Telecommimications” Commission, “The European Conference of Posts and Telecommunications Administrations, consider iig - that users and suppliers of Telecommunications services are interested in the quality of service irrespective of how that service is provided, - that Quality of Service
3、 (QOS) is a statement of the performance of all aspects of the service offered to the user, and that proposals for a general QOS framework are described in CEPT Recommendation T/SF 54 E and CCITT Recommendation G. 106, - that Network Performance (NP) describes the externally perceived performance at
4、tributes of the network and is independent of terminal performance and user actions, - that CCITT Recommendations X.134 through X.137 define the NP parameters for packet switched services, - that no implementation of measurement methods and tools for these NP parameters is yet defined, - that there
5、is an urgent need for ready-to-implement methods and tools for evaluating the main aspects of the O network performance part of the international packet switched service, recommends that the following set of performance indicators is used by CEPT Administrations in measuring, evaluating and communic
6、ating the network performance of the packet switched service.” 1. INTRODUCTION In order to evaluate the network performance it is necessary to have information on the most important elements of the general performance criteria: availability, dependability, accuracy and speed of service. Indicators p
7、roviding this information should, where possible, be derived from information sources, records and events available in most networks. Also the indicators should be easy to derive and, as far as possible, related to existing NP parameters. 2. INDICATORS For each of the above performance criteria indi
8、cators are chosen which are representative of the service. Availability Dependability Accuracy - see Note Speed of service - Unsuccessful NC calis ratio (UNCR) - Mean time between NC disconnections (MTNC) - Mean time between Resets (MTRS) - Transmitted throughput (TTP) - Received throughput (RTP) -
9、Round-trip delay (RTD) - Call set-up delay (CSD) Note. No indicators are proposed for the accuracy aspect of service. Under normal conditions the accuracy performance of a PSPDN cannot be measured in a simple but representative manner. Edition of January 8, 1990 CEPT T/CAC*Z*E 89 = 232b414 0002907 4
10、 T/CAC 2 E Page 2 3. RELATIONSHIP WITH CCITT NETWORK PERFORMANCE PARAMETERS UNCR is related to the complement of the X. 137 parameter Service availability. MTRS is related to the inverse ot the X.136 parameter Reset probability. MTNC is related to the inverse of the X. 136 parameter Premature discon
11、nect probability. TTP and RTP are related to the X.135 parameter Throughput capacity. RTD is related to twice the X.135 parameter Data packet transfer delay. CSD is related to the X.135 parameter Call set-up delay. 4. PRIORITIES Of the above indicators UNCR, MTNC, TP and RTP are considered to be ess
12、ential for a fist investigation of the performance of PSPDNs. 5. IMPLEMENTATION A distinction is made between indicators which can be derived ?internally?, using, for example, existing billing or statistical tools, and indicators which have to be determined ?externally? requiring dedicated test equi
13、pment and procedures. The indicators derived internally are: - Unsuccessftil NC calls ratio (UNCR) - Mean time between NC disconnections (MTNC) - Mean time between resets (MTRS) The indicators derived externally are: - Transmitted throughput (TTP) - Received throughput (RTP) - Round-tsip delay (RTD) - Call set-up delay (CSD) Definitions, requirements, procedures and processing for each of the internally derived indicators are described in Recommendation T/CAC 3 E. The externally derived indicators are similarly described in Recommendation T/CAC 4 E. Edition of January 8, y90 ?
