ANSI ATIS T1.TR.42-1995 Enhanced Analysis of FCC - Reportable Service Outage Data.pdf

1、Report No. 42 A Technical ReportonAugust 1995 Enhanced Analysis ofFCC-Reportable ServiceOutage DataPrepared byT1A1.2Working Group onNetwork SurvivabilityPerformanceCommittee T1 is sponsored by the Alliance for Telecommunications Industry Solutions(formerly the Exchange Carriers Standards Association

2、)Accredited by American National Standards InstituteCopyright 1995 by Alliance for Telecommunications IndustrySolutions All rights reserved.No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission of the publisher

3、.Technical Report on Enhanced Analysis ofFCC-Reportable Service Outage DataAbstractIn response to a Network Reliability Steering Committee (NRSC) request,this Technical Report (TR) enhances the methods (guidelines andalgorithms) given in Committee T1 TR No. 38 for analyzingFCC-reportable service out

4、age data to determine the impact of serviceoutages. This TR replaces TR No. 38, and is also being considered by theT1A1.2 Working Group for the development of a longer term TR onNetwork Survivability Performance.Document T1A1.2/95-001R4Prepared by T1A1.2Working Group on Network Survivability Perform

5、anceForewordThis Technical Report (TR) addresses concerns from thetelecommunications community about the survivability and reliability oftelecommunications networks, including the services provided. It isintended to provide a basis for analyzing service outage data that isreported to the Federal Com

6、munications Commission (FCC) anddetermining their impact.This TR provides enhanced methods (guidelines and algorithms) comparedto those contained in Committee T1 TR No. 38 that it replaces. It is alsobeing considered by the T1A1.2 Working Group for the development of alonger term TR on Network Survi

7、vability Performance.The intended audience of this report includes telecommunications networkand service regulators, providers and users of telecommunicationsnetworks and services.Suggestions for improvement of this report are welcome. They should besent to the Alliance for Telecommunications Indust

8、ry Solutions, Suite 500,1200 G Street, N.W., Washington, D.C. 20005.Working Group T1A1.2 on Network Survivability, which developed thisreport, had the following officers and active participants:Chair: F. KaudelVice-Chair: S. MakrisEditors:F. KaudelH. DaughertyS. MakrisM. PeatsActive Participants:J.

9、BennettM. BruscaR. CanadayP. ChristensenM. DaneshmandH. DaughertyS. DhawanR. DoverspikeC. EcholsJ. HealyH. HoltonR. JensenF. KaudelM. KocaturkJ. LordS. MakrisS. NathanM. PeatsA. RauschA. ReillyJ. RoquetM. WarrenTechnical Report No. 42iTable of Contents0. Executive Summary 11. Purpose, Scope, Applica

10、tion, and Outline 21.1 Purpose 21.2 Scope. 21.3 Application. 21.4 Outline. 22. Related Work . 22.1 Committee T1 Standards Work .2.2 International Standards Work 32.3 Other Domestic Forums and Committees3. Introduction . 44. Desirable Outage Index Properties 55. Definition of Outage Index5.1 Data Use

11、d . 55.1.1 FCC Data Items .5.1.2 NRSC Outage Categories.5.2 Overall Definition 65.3 Service Weights.5.4 Duration Weights. 75.5 Magnitude Weights. 85.5.1 Outage Types.5.5.2 Outage Categories.5.5.3 Calculation of Magnitude Weights . 95.5.4 Lines Method. 125.5.5 Blocked Calls Method 126. Examples of Ou

12、tage Index Calculation 136.1 Dedicated (Local Switch) Outage Example. 136.1.1 Report Data . 136.1.2 Outage Index Calculation6.2 Diversified (Tandem Switch) Outage Example 146.2.1 Report Data . 146.2.2 Outage Index Calculation 146.3 Diversified (Facilities) Outage Example 156.3.1 Report Data . 156.3.

13、2 Outage Index Calculation7. Aggregated Outage Index 8. Application to FCC-Reportable Service Outage Data. 169. Future Work 1710. Summary 1711. Definitions 1812. Acronyms 1913. Bibliography 20Appendix A. Analysis of the Number of Outages 21A.1 Analysis of the Number of Outages with Respect to a Refe

14、rence Period 21Technical Report No. 42iiA.2 Trend Analysis. 22A.2.1 Statistical Theory 23A.2.1.1 Poisson Distribution . 23A.2.1.2 Testing for Evidence of a Trend 24A.2.1.3 Calculating the Parameters for the Model. 25A.2.1.4 Predicting Future Outage Rates 26A.2.2 Example Test for Evidence of a Trend

15、Using a 95% Level of Confidence . 26A.2.3 Trend Analysis of FCC-Reportable Service Outage Data (Number of Outages) 28Appendix B. Additional Examples of Outage Index Calculation . 30B.1 Dedicated (Local Switch) Partial Services Outage Example with Same Service OutageDurations 30B.2 Dedicated (Local S

16、witch) Partial Services Outage Example with Different Service OutageDurations 31B.3 Diversified (Tandem Switch) Outage Example 32B.4 Diversified (Facilities, DCS) Outage Example . 33B.5 Diversified (Facilities, Fiber) Outage Example. 34Appendix C. Gradual Restoration of Service. 35C.1 Example.35C.2

17、General Technique 36Appendix D. Estimating Confidence Intervals for Aggregated Outage Index Values Based on Moments ofthe Index Distribution 37D.1 Bootstrap Techniques 37D.2 Closed-Form Techniques 38D.2.1 TheoremD.2.2 Application to Outage Index Using FCC-Reportable Service Outage Data 38D.2.2.1 Est

18、imate the Mean and Standard Deviation of the Annual Sum ofOutage Index Values Based on the Baseline Year . 38D.2.2.2 Estimate the Mean and Standard Deviation of the Quarterly Sum ofOutage Index Values Based on the Baseline Year . 38D.2.3 Uses of the Mean and Standard Deviation Estimates. 39D.3 Power

19、 of the Test. 39Appendix E. Normalized Outage Index 41E.1 Normalization in the ATIS NRSC 1994 Annual Report 41E.1.1 Normalization Data .E.1.2 Normalization of Magnitude 41E.1.3 Normalization of Frequency E.1.4 Normalization Factors 42E.1.5 Calculation Based on Normalized Customers 42E.1.5.1 Frequenc

20、y Normalization 42E.1.5.2 Index Normalization E.2 Generalization for Future Reports 43E.2.1 Normalization Factors Example. 43E.2.2 Frequency NormalizationE.2.3 Index Normalization 44E.2.4 Analysis of Normalized Data. 44Index 45Technical Report No. 4210 . Executive SummaryThis Technical Report (TR) e

21、nhances the methods (guidelines and algorithms) contained inCommittee T1 Technical Report No. 38 “A Technical Report on Analysis of FCC-ReportableService Outage Data,” August 1994. The purpose of this TR is to assist the telecommunicationsindustry and users in analyzing CC Docket 91-273 service outa

22、ge data that is reported to theFederal Communications Commission (FCC). In particular, this report provides a response to aNetwork Reliability Steering Committee (NRSC) request for enhancements to TR No. 38. TheNRSC is a committee of the Alliance for Telecommunications Industry Solutions (ATIS, form

23、erlythe Exchange Carriers Standards Association ECSA) with the mission to ensure a continued highlevel of network reliability.The following list summarizes the major changes made to TR No. 38 in developing thisreplacement TR: replaced tables for Duration Weight and Magnitude Weight with equations, m

24、ade modifications to reflect changes in FCCs reporting requirements (i.e., only the“real-time blocked calls” in diversified outages are divided by 3), revised time factors including introduction of a “Weekend” Time Factor, introduced procedures to calculate the outage index when services are gradual

25、ly restored, developed a technique for estimating confidence intervals for aggregated outage indexvalues, and documented the normalization process used in the ATIS NRSC 1994 Annual Report, andproposed a more general normalization process for comparative and trend analysisapplications.The methods in

26、this TR yield an outage index based on a combination of service(s) affected,duration, and magnitude (number of customers affected). Various properties are believed to beessential for such an index. In particular, the outage index should: reflect the relative importance of outages for different servi

27、ces, be able to be aggregated to allow comparisons over time, and reflect the impact of small and large outages as they are perceived by the public.Calculation methods and service outage examples including outages caused by the failure of alocal switch, a tandem switch, and facilities are provided,

28、based on both FCC-reportable serviceoutage data and NRSC outage categories. These methods apply to both past and future serviceoutage data. In addition to the definition of the outage index, this TR provides a trend analysis foroutage counts, additional examples of outage index calculations (e.g., a

29、ccounting for gradualrestoration of services), calculations for confidence intervals, and normalization processes.Technical Report No. 4221 . Purpose, Scope, Application, and Outline1 . 1 Purpose This Technical Report (TR) provides methods (guidelines andalgorithms) to the telecommunications industr

30、y and users for analyzing CC Docket 91-273 1,2service outage data that is reported to the Federal Communications Commission (FCC). Theindustry is expected to use these methods to better analyze network outages. In particular, this TRprovides a response to a Network Reliability Steering Committee (NR

31、SC) request for input to itsquarterly and annual reports to the FCC and the telecommunications industry. The NRSC is acommittee of the Alliance for Telecommunications Industry Solutions (ATIS) with the mission toensure a continued high level of network reliability.1 . 2 Scope This TR enhances and re

32、places the work of TR No. 38 3.1 . 3 Application This TR will enable the telecommunications industry and users toevaluate the reliability and quality of services in telecommunications networks.1 . 4 Outline Section 2 reviews related work. Section 3 is the introduction to this TR.Section 4 describes

33、general, desirable properties of an outage index. Section 5 defines the outageindex recommended in this TR, by specifying the FCC and NRSC data used as well as thealgorithms for calculating the outage index. Section 6 gives three detailed outage examples of howto apply the algorithms given in Sectio

34、n 5 (for local switch, tandem switch, and facilities failurecases). Sections 7 and 8 discuss calculation of an aggregate index for multiple outages andapplication to FCC-reportable service outage data. Section 9 suggests areas for future work, andSection 10 gives a summary. Sections 11, 12, and 13 p

35、rovide definitions, acronyms, and abibliography, respectively. Appendix A provides two techniques for analyzing the number ofoutages in a time period, based on FCC-reportable service outage data. Appendix B givesadditional examples of outage index calculations. Appendix C provides procedures for cal

36、culatingthe outage index in cases of gradual restoration of service(s). Appendix D furnishes equations forestimating confidence intervals for aggregated outage index values. Appendix E gives processes tonormalize the outage index for differences in time periods.2 . Related Work2 . 1 Committee T1 Sta

37、ndards Work In addition to T1A1, the following three otherT1 technical subcommittees are involved in work related to telecommunications networksurvivability performance: T1E1, T1M1, and T1S1.T1E1 is considering environmental standards (e.g., operating conditions, earthquakes, and fires).A TR is bein

38、g developed for enhancing network survivability during and after power failures. It isexpected to identify and document possible approaches to the reduction of battery current drainduring commercial power or rectifier loss, and to the orderly removal and restoral of redundant,nonessential, or less i

39、mportant equipment to prolong essential service.T1M1 is considering the Operations, Administration and Maintenance (OA thecustomer threshold was subsequently reduced to 30,000 7. In addition, the FCC determined thatnetwork outages affecting special facilities must be reported, regardless of the abov

40、e thresholds7. The FCC defines as a special facility “a major military installation, a key government facility,a nuclear power plant or a 911 tandem, or a large or medium commercial service airport, asdefined and listed in the National Plan of Integrated Airport Systems.” 7.In an effort to address t

41、he FCC challenge, methods to quantify outages are presented in this TR.The methods consist of guidelines and algorithms to compute an outage index, based on FCC dataitems and NRSC outage categories, only for those outages meeting the 30,000 potentially affectedcustomers and 30-minute duration criter

42、ia.The following list summarizes the major changes made to TR No. 38 in developing thisreplacement TR: uses equations, instead of tables, for the Duration Weight and the Magnitude Weight, makes a distinction between “real-time blocked calls” and “historical carried call count” indiversified outages,

43、 estimates customers affected as “real-time blocked calls” divided by 3 in diversifiedoutages, as per FCC Memorandum Opinion and Order and Further Notice of ProposedRulemaking, 93-491 8, gives power functions of the test for quarterly outage frequency for the FCC-reportableservice outages, revises t

44、he time factors for the “Evening” and “Night” periods and introduces a Time Factorfor “Weekend” periods, revises the Service Weight table by rearranging the columns and by including the “MethodUsed,” introduces a procedure to calculate the outage index for outages with gradual restoration ofservice(

45、s), presents two techniques for estimating confidence intervals for aggregated outage indexvalues, describes the normalization process used in the ATIS NRSC 1994 Annual Report 9, witha generalization for future comparative and trend analysis applications, and provides examples of outage index calcul

46、ations for combined service outages.Technical Report No. 4254 . Desirable Outage Index PropertiesAn outage index (I) is a function which assigns a numerical value I(O) to an outage O. Thefunction is designed as a measure of the customer impact which integrates three “data items” Services Affected, D

47、uration, and Magnitude (number of customers affected) into an index.Since every outage has some impact (although negligible in many cases), a useful intuitive propertyfor an index is that it is non-negative (i.e., I(O) 0 for any outage O). Another intuitive propertyis that I(O1) I(O2) if outage O1is

48、 considered more severe than outage O2. Below are threeadditional properties believed to be essential for this index function:Property 1: The index should account for the relative importance of the different servicesaffected by the outage. In other words, if two outages O1and O2have the same duratio

49、nand magnitude, but each affects a different service, S1and S2, then I(O1) I(O2) if S1isconsidered a more important service than S2.Property 2: Aggregated index values calculated by summing individual outage index valuesover time periods should allow comparisons across time periods.Property 3: The aggregate outage index for a year with several small outages is less than theindex for a year with one outage where its duration (or magnitude) equals the sum of thedurations of the small outages. On the other hand, the aggregate outage in