tpcc_current (Дополнительные методические материалы), страница 2

857.1Pricing Method ology....................................................................................................................... 857.2Priced System ................................................................................................................................... 857.3Requ ired Rep orting .........................................................................................................................

88Clau se 8: FULL DISCLOSURE ................................................................................................................................ 898.1Fu ll Disclosu re Rep ort Requ irem ents ........................................................................................... 898.3Revisions to the Fu ll Disclosu re Rep ort ....................................................................................... 98Clau se 9: AUDIT ..................................................................................................................................................... 1009.1General Ru les ................................................................................................................................. 1009.2Au d itor's check list ........................................................................................................................

100Ind ex105Ap p end ix A: SAMPLE PROGRAMS ................................................................................................................... 108A.1The N ew -Ord er Transaction ........................................................................................................ 108A.2The Paym ent Transaction ............................................................................................................. 110A.3The Ord er-Statu s Transaction ...................................................................................................... 112A.4The Delivery Transaction .............................................................................................................

114A.5The Stock-Level Transaction ........................................................................................................ 116A.6Sam p le Load Program .................................................................................................................. 117Ap p end ix B: EXECUTIVE SUMMARY STATEMEN T ...................................................................................... 130Ap p end ix C: N UMERICAL QUAN TITIES SUMMARY ................................................................................... 132TPC Benchm ark™C - Stand ard Sp ecification, Revision 5.11 - Page 6 of 130Clause 0: PREAMBLE0.1IntroductionTPC Benchm ark™C (TPC-C) is an OLTP w orkload .

It is a m ixtu re of read -only and u p d ate intensive transactionsthat sim u late the activities fou nd in com p lex OLTP ap p lication environm ents. It d oes so by exercising a bread th ofsystem com p onents associated w ith su ch environm ents, w hich are characterized by:• The sim u ltaneou s execu tion of m u ltip le transaction typ es that sp an a bread th of com p lexity• On-line and d eferred transaction execu tion m od es• Mu ltip le on -line term inal sessions• Mod erate system and ap p lication execu tion tim e• Significant d isk inp u t/ ou tp u t• Transaction integrity (ACID p rop erties)• N on-u niform d istribu tion of d ata access throu gh p rim ary and second ary keys• Databases consisting of m any tables w ith a w id e variety of sizes, attribu tes, and relat ionship s• Contention on d ata access and u p d ateThe p erform ance m etric rep orted by TPC-C is a "bu siness throu ghp u t" m easu ring the nu m ber of ord ers p rocessedp er m inu te.

Mu ltip le transactions are u sed to sim u late the bu siness activity of p rocessing an ord er, and eachtransaction is su bject to a resp onse tim e constraint. The p erform ance m etric for this benchm ark is exp ressed intransactions-p er-m inu te-C (tp m C). To be com p liant w ith the TPC-C stand ard , all references to TPC-C resu lts m u stinclu d e the tp m C rate, the associated p rice-p er-tp m C, and the availability d ate of the p riced configu ration.To be com p liant w ith the op tional TPC-Energy stand ard , the ad d itional p rim ary m etric, exp ressed as w atts-p ertp m C m u st be rep orted .

The requ irem ents of the TPC-Energy Sp ecification can be fou nd at w w w .tp c.org.Althou gh these sp ecifications exp ress im p lem entation in term s of a relational d ata m od el w ith conventional lockingschem e, the d atabase m ay be im p lem ented u sing any com m ercially available d atabase m anagem ent system (DBMS),d atabase server, file system , or other d ata rep ository that p rovid es a fu nction ally equ ivalent im p lem entation. Theterm s "table", "row ", and "colu m n" are u sed in this d ocu m ent only as exam p les of logical d ata stru ctu res.TPC-C u ses term inology and m etrics that are sim ilar to other benchm arks, originated by the TPC or others.

Su chsim ilarity in term inology d oes not in any w ay im p ly that TPC -C resu lts are com p arable to oth er benchm arks. Theonly benchm ark resu lts com p arable to TPC-C are other TPC-C resu lts conform ant w ith the sam e revision.Desp ite the fact that this benchm ark offers a rich environm ent that em u lates m any OLTP ap p lication s, thisbenchm ark d oes not reflect the entire range of OLTP requ irem ents. In ad d ition, the extent to w hich a cu stom er canachieve the resu lts rep orted by a vend or is highly d ep end ent on how closely TPC -C ap p roxim ates the cu stom erap p lication.

The relative p erform ance of sy stem s d erived from this benchm ark d oes not necessarily hold for otherw orkload s or environm ents. Extrap olations to any other environm ent are not recom m end ed .TPC Benchm ark™C - Stand ard Sp ecification, Revision 5.11 - Page 7 of 130Benchm ark resu lts are highly d ep end ent u p on w orkload , sp ecific ap p lication requ irem ents, and system s d esign andim p lem entation. Relative system p erform ance w ill vary as a resu lt of these and other factors. Therefore, TPC -Cshou ld not be u sed as a su bstitu te for a sp ecific cu stom er ap p lication benchm arking w hen critical cap acity p lan ningand / or p rod u ct evalu ation d ecisions are contem p lated .Benchm ark sp onsors are p erm itted several p ossible system d esigns, insofar as they ad here to the m od el d escribedand p ictorially illu strated in Clau se 6.

A Fu ll Disclosu re Rep ort of the im p lem entation d etails, as sp ecified in Clau se8, m u st be m ad e available along w ith the rep orted resu lts.Comment: While sep arated from the m ain text for read ability, com m ents are a p art of the stand ard and m u st beenforced . H ow ever, the sam p le p rogram s, inclu d ed as Ap p end ix A, the su m m ary statem ents, inclu d ed as Ap p end ixB, and the nu m erical qu antities su m m ary, inclu d ed as Ap p end ix C, are p rovid ed only as exam p les and aresp ecifically not p art of this stand ard .0.2General Implementation GuidelinesThe p u rp ose of TPC benchm arks is to p rovid e relevant, objective p erform ance d ata to ind u stry u sers.

To achievethat p u rp ose, TPC benchm ark sp ecifications requ ire that benchm ark tests be im p lem ented w ith system s, p rod u cts,technologies and p ricing that:• Are generally available to u sers.• Are relevant to the m arket segm ent that the ind ivid u al TPC benchm ark m od els or rep resents (e.g. TPC -Am od els and rep resents high -volu m e, sim p le OLTP environm ents).• A significant nu m ber of u sers in the m arket segm ent the benchm ark m od els or rep resents w ou ld p lau siblyim p lem ent.The u se of new system s, p rod u cts, technologies (hard w are or softw are) and p ricing is encou raged so long as theym eet the requ irem ents above. Sp ecifically p rohibited are benchm ark system s, p rod u cts, technologies, p ricing(hereafter referred to as "im p lem entations") w hose p rim ary p u rp ose is p erform ance op tim ization of TPC benchm arkresu lts w ithou t any corresp ond ing ap p licability to real-w orld ap p lication s and environm ents.

In other w ord s, all"benchm ark sp ecials," im p lem entations that im p rove benchm ark resu lts bu t not real-w orld p erform ance or p ricing,are p rohibited .The follow ing characteristics shou ld be u sed as a gu id e to ju d ge w hether a p articu lar im p lem entation is abenchm ark sp ecial. It is not requ ired that each p oint below be m et, bu t that the cu m u lative w eight of the evid encebe consid ered to id entify an u naccep table im p lem entation.