Introduction to tablebase and tablebase Family. William Weber. Market Experts Distribution, SL

Size: px

Start display at page:

Download "Introduction to tablebase and tablebase Family. William Weber. Market Experts Distribution, SL"

Agatha O’Neal’
5 years ago
Views:

1 Introduction to tablebase and tablebase Family William Weber Market Experts Distribution, SL

2 Presentation Outline Introduction to DataKinetics tablebase tablebase Product Family

3 Who Is DataKinetics? Established in 1977 The global leader in data performance and optimization solutions. Solving IT problems and reducing IT costs for the Fortune 500 IBM Business partner Founder of The Planet Mainframe Alliance

4 What Does DataKinetics Do? Our comprehensive, world-renowned suite of solutions enables Fortune 500 companies to: Process over a billion mission-critical transactions per second every day Accelerate application processing by up to 98% Seamlessly integrate data on mainframe and distributed systems Enable big data and analytics in multi-platform environments Enable increased control and flexibility in sub-capacity pricing and R4H soft capping environments

global banks 3 of the top 5 global credit card companies 3 of the top

5 Who Does DataKinetics Work With? Some of the world s largest companies are our clients: 3 of the top 5 global banks 3 of the top 5 global credit card companies 3 of the top 5 US property & casualty insurance companies 2 of the top 5 US health insurance companies 20% of the Fortune 50

6 Presentation Outline Introduction to DataKinetics tablebase tablebase Product Family

7 What is tablebase? tablebase is an in-memory accelerator Define, build, maintain and manage in-memory tables Phenomenally fast access to data in the tables Optimize table search Decrease elapsed time It s just physics, math and great algorithms Take data that is highly accessed Load it in memory close to the application Repeatedly access this data using the shortest and fastest path Decrease CPU Improve Response time

8 How fast? IBM Benchmark Results for DB2 Both results use data that is in-memory Only one uses the shortest fastest path When the amount of time taken for each transaction is reduced, and the number of transactions per unit of time is increased, performance and capacity are increased significantly

9 Shortest path to data

10 One customer s experience Challenge A COBOL program was using an internal table and a binary search The search code was called 1.25 million times and had 4 searches in it Took over an hour of CPU to execute Solution Replace the 4 searches with calls to tablebase Results 99.98% reduction in CPU seconds Now takes less than a minute to execute

Results From Credit Card Processing Challenge Reconciliation batch processing taking too long Solution Move a table describing the credit card options

11 Results From Credit Card Processing Challenge Reconciliation batch processing taking too long Solution Move a table describing the credit card options into tablebase Each transaction required data from that table Results 97% reduction in elapsed time Batch job that took 8 hours to complete now takes 15 min

12 Uses for tablebase Replacing CICS tables and TS queues with tablebase tables Replace hard coded module tables with tablebase tables Replace VSAM (KSDS and RRDS) files with tablebase tables Unload high referenced DB2 table data into tablebase tables Replace BDAM files with tablebase tables Use tablebase tables for summarizing and grouping data for reports Use tablebase for DB2 cursor scrolling tablebase is used for storing cursor results for DB2 cursor scrolling in CICS Replace COBOL SEARCH with tablebase search module tablebase can be used to provide session continuation data for online applications Replace I/O sort with a virtual sort using alternate indices that can be defined in real time Assemble data from multiple sources into a temporary table for subsequent processing or renderings Place reference data and business rules in read-only tables Replace programmatic logic trees with externalized decision tables

13 Discovery DB2 Apptune, Detector GETPAGES GETPAGES SYNC I/O SYNC I/O READ NON-READ OBJECT # OF ROW TABLE NAME COUNT TIME COUNT TIME EXECUTNS EXECUTNS TYPE ROWS LENGTH GB PITBPREF 139,102,000 12: : ,327,028 0T T_DELIVERY_CHANNEL 72,922,988 07: : ,195,936 0T PROCESS_LOCATION 39,691,687 03: : ,434,286 0T REGIONS 39,564,353 03: : ,376,234 0T CUSTOM_PHOTO_RULES1 31,777,736 02: : ,888,868 0T CALL_TIMES_TB 21,738,834 02: : ,037,498 0T PITBSPRI 9,633,519 00: :00.1 6,422,346 0T SLCTN_PREFN_DFNTN 7,361,714 00: :00.1 4,907,787 0T PITBPRDC 15,389,568 01: :01.2 3,080,381 0T PROD_CORP 18,696,038 00: :06.4 1,696,716 0T TEST_TB 5,205,402 00: :00.3 1,671,653 0T ACCT_RELCODE_PROCESS 2,833,120 00: :06.2 1,416,667 0T NBTBTRN 1,301,645 00: : ,822 0T CARD_STATUS_TB 10,275,312 00: : ,848 0T PITBCHGC 2,949,616 00: : ,925 0T

14 DB2 Stats Query Collect Stats on Multiple Dates Process to find read/write rates DB NAME TS NAME SCHEMA TBL NAME UPD TIMESTAMP #ACCESSES #WRITES SIZE(KB) PRTS D0350C0C SBANC098 C0C MS_BUFFER_REGISTRY D0350C0C SBANC490 C0C BATCH_JOB_EXECUTIO D0387C1B RTS9048 C1B RULEPACKAGE D0388C2A TSTFTM01 C2A OBJ_BASE D1419A65 S1419 DA65 T D1484D2W S4053 DB2W T7997_CTL_ACT_REMT D1484D2W SF721 DB2W TE111_AMX_INI_CM_P D1558AP1 S5962 APPC1 T5962_FS_ALIAS_ASC E D1961A65 S1961 DA65 T D2075A65 S2075 DA65 T D2120M0B S2120 DB2W T2120_SEAS_RECOMM D2138M0B S2138 DB2W T2138_UIR_ROUTING D2139M0B S2139 DB2W T2139_UIR_TRANS Write Write Write D2141M0B S2141 DB2W T2141_UIR_HISTORY Rate Time Read Rate Rate Time Read Rate D2782A66 S7467 DA66 T7467_MER_FULFIL_R D2800A65 S2800 DA65 T D2873A65 S2873 DA65 T D3100MUS S3183 MUS T3183_ALRT_FILE_PT D3100MUS S3197 MUS T3197_EVENT_SUB_PO D3106A66 S7791 DA66 T D3643C01 PR4RULEN C01 PR4_RULE D3643C02 PR4RULEN C02 PR4_RULE D3643CB9 PGTS0080 MB9E1 PR_SYS_RULESET_IND D3643CB9 PGTSP034 MB9E1 PR_ASSEMBLEDCLASSE March 3 to March 8 March 8 to March 11 March 11 to March 31 March 3 to March 31 Write UID Time Read Rate Rate Time Read Rate Rate DEODE102.PGTS0070.PEORE102.PR4_RULE_VW D3986MUS.S3986A.MUS.T D7099MTC.SH627.MTC.TF465_WRKR_CRT_SIG DD927MK3.SD927.MK3.TD927_PTY_NM E D3643MBA.PGTSP039.MBA.PR_SYS_CACHE_DEP D2E50M0B.S2E50.DB2W.T2E50_COMMUN_DATA D2075A65.S2075.DA65.T D1419A65.S1419.DA65.T DEODE102.PGTS0170.PEODE102.PR_SYS_STATUSDETAI D4259MUS.SZUPT.MUS.TA985_CA_SMRY

15 Discovery VSAM SMF 64 Records Data Set name EXCPS INSERTS DELETES UPDATES RETRIEVALS RECORDS CHANGE # RECORDS LRECL GB PRDFE.#LPS.APOR.DATA 24,883,173 1, , ,193, , PRDI9.#P1BHI.N1ZXFDR.DATA 28,821, ,137, , PRDXW.#ENTRDX.VPF.DK 2,319,082 95, , ,925,834-68, , PRDI9.#P1BHI.PE01RCR.X00001.DATA 26,092,817 10, ,958,864 10,559 2,952, PRDI9.#P1BHI.PG15IDF.X00021.DATA 32,675,941 97,947 73, ,488,668 24,216 6,452, PRDI9.#P1BHI.E10FRPM.A00015.DATA 3,180, ,269, , PRDI9.#P1BHI.E104CIF.A00004.DATA 5,942, ,679, , PRDI9.#P1BHI.N1ZXFDRX.DATA 24,330,441 17,670 19, ,776,087-2,026 4,364, PRDI9.#P1BHI.E103CIF.A00003.DATA 4,433, ,130, , PRDI9.#P1BHI.N105AFS.A00005.DATA 12,132, ,528, , PRDE6.OS.PAYEE.INDEX.CICS.DATA 79, ,876, , PRDI9.#P1BHI.PG0WAIF.X00032.DATA 11,813,441 86, ,542,171 86,676 2,882, PRDI9.#P1BHI.N113PAC.A00039.DATA 4,467, ,373, , PRDI9.#P1BHI.PD01RCR.A00001.DATA 9,576, ,305, , PRDI9.#P1BHI.E10FRPM.A00015.DATA 496, ,503, ,

16 Strobe (or other Application Profiling Tool) Analyze Programs See specific SQL/VSAM CPU/wait times Determine specific SQL statements and program modules Measure tablebase Specific Improvements Compuware Strobe Specific Incremental reports specific for finding tablebase Candidates

17 Table Types Defining a Table As well as Organization and Search Methods, tables are made up of Rows, Keys, Index. In-memory tables are kept in a Table Space Region.

18 Multiple Organizations/Retrieval Organizations Sequential (asc, desc) Hash Random User Defined Sequence Retrieval Methods Serial Binary Queued Sequential Bounded Binary Hash Creating the index Using the index

19 Presentation Outline Introduction to DataKinetics tablebase tablebase Product Family

20 DataKinetics tablebase Product Family tablebase VTS Shared in-memory tables enabling parallel access to data VTS Manager Extends tables sharing across LPARS and 24x7x365 operation VTS Gate Additional security for memory resident VTS VTSEdge ziip eligble VTS API with JS integration engine enabling CPU effective, remote, and flexible APIs to in-memory data SQLKinetics Leveraging tablebase without code changes (for DB2)

21 Q&A

22 Thank you for your time. William Weber Market Experts Distribution, SL Tel:

Case Study. British Library 19th Century Book Digitisation Project

Case Study. British Library 19th Century Book Digitisation Project Case Study British Library 19th Century Book Digitisation Project I. Introduction 1. About the British Library The British Library is the national library of the United Kingdom. It holds over 150 million