* copyright (C) 1984-2019 merrill consultants dallas texas usa

but then you could not go back to determine the Primary

names, if they were ever needed. Thus there was no code

changed in MXG by this Documentation Note Change.

Thanks to Joan Nielsen, PKS Information Services, Inc, USA.

====Changes thru 14.271 were included in MXG 14.08 dated Nov 13, 1996===
Change 14.271 Support for AS/400-OS/400 Release 3.7.0 added many new

VMACQAPM variables to datasets QAPMBUS,QAPMCONF,QAPMJOBS,QAPMMIOP,

Nov 12, 1996 and QAPMRESP. Support for all fields added by 3.6.0 are

also now supported. All 3.7.0 changes were compatible,

except for the QAPMETH dataset, which had ETLFRT expanded

from three to six bytes (and I had misspelled it!). The

3.6.0 partial changes in CHANGE 14.098 were INCOMPATIBLE.

Thanks to Clark Jennings, Reynolds Metals Company, USA.

Change 14.270 Cosmetic. All of the LABEL, FORMAT, INFORMAT, and LENGTH

VMAC110 statements in the Statistics code (Subtype=2) were moved

Nov 12, 1996 together into single statements with alphabetic lists.
Change 14.269 Execution of MXG Software under Unix requires each PDS

IEBUPDTE member become an individual file with suffix "sas", AND

Nov 12, 1996 the filename must be in lower case! The IEBUPDTE.SAS

program creates the individual files, but their names are

in upper case, causing execution under UNIX to fail!

This change inserts MEMBER=LOWCASE(MEMBER); after the

MEMBER=TRIM.... statement to create the members in lower

case. Make sure the output directory is empty before

running the IEBUPDTE.SAS program, as it appears that if

the output file name already exists as an upper case name

that the data is copied but the name is not re-cased!

Thanks to Mike Orlando, Nalco Chemical Company, USA.

Change 14.268 Protection was added to eliminate DIVISION BY ZERO error

VMAC7072 if IOCCOEFF was zero. I had assumed IOCCOEFF would never

Nov 12, 1996 be zero, but if it was, the divide error occurred.

RMFINTRV Also, PGPIOTM is now forced missing for MVS/ESA 5.2 or

Mar 10, 1997 higher. When IO units could be based on either EXCP or

IOTM, both PGPEXCP and PGPIOTM (raw IO EXCP count or raw

IO Connect time) were calculated (even though only one

was ever valid), because IBM did not flag whether you had

specified EXCP or IOTM for your IO Service Units). But

with MVS/ESA 5.2+, IO Service Units can only be based on

EXCP count, so PGPIOTM has no meaning.

Mar 10, 1997: Because PGPIOTM in TYPE72 is missing, the

variables BATIOTM, CICSIOTM, IMSIOTM, TSOIOTM, OTHRIOTM

in dataset RMFINTRV will also be missing.

Change 14.267 This report showed zeros for channel busy in LPAR mode.

ANALPATH Change PNCHANBY=MIN(PNCHANBY,PCHANBY);

Nov 12, 1996 to read PNCHANBY=MAX(PNCHANBY,PCHANBY);

Nov 18, 1996 (Nov 12 text had PCHANBY on left, corrected Nov 18.)

Thanks to Jim Ray, Branch Banking and Trust, USA.
Change 14.266 MXG 14.04-MXG 14.07. The word "NEW" must be deleted from

VMACITRF line 245 to prevent INPUT STATEMENT EXCEEDED ERROR. This

Nov 4, 1996 was introduced by Change 14.128 and incomplete testing.

Thanks to Siegfried Trantes, IDG Informationsvararbeitung, GERMANY.

Change 14.265 Support for HP's MeasureWare for SUN operating system.

EXHPSUTT The same datasets that were created by HP's PCS for SUN

IMACMWSU are created, with one new dataset:

TYPEMWSU Dataset Description

VMACMWSU HPSUTRAN Transaction Tracker Record

Nov 1, 1996

Thanks to Gary Alexander, BMC, USA.
Change 14.264 Support for CA's ENDEVOR SMF record creates two datasets

EXENDVAC Dataset Description

EXENDVSE ENDEAVAC Action Record

FORMATS ENDEAVSE Security Record

IMACENDV The Action record has been validated with actual data,

TYPEENDV but no security records had been created (there were no

VMACENDV security violations!) so that code is untested.

Nov 1, 1996

Thanks to Grace Pergament, Amdahl Corporation, USA.

Thanks to Susan Walters, Michelin Tire Corporation, USA.

Change 14.263 Dataset DCOLAG (Aggregate Group Definitions) has several

VMACDCOL fields incorrect, because the @208 preceding DAGEXPYR

Nov 1, 1996 should have been @209.

Thanks to Gary Miner, Fidelity Investments, USA.

Thanks to Lawrence Jermyn, Fidelity Investments, USA.
Change 14.262 Calculation of DOMDSPCH and ADJTARG must be changed to:

VMACAPAF IF ACTIVE GT 0 THEN DO;

Nov 1, 1996 DOMDSPCH=INTERVAL*DOMDSPCH/ACTIVE;

ADJTARG =INTERVAL*ADJTARG /ACTIVE;

END;

(The variables are divided by INTERVAL earlier in code.)

Thanks to Stephen H. Lothes, Flexlogic, Inc, USA.
Change 14.261 Using the new High Level Assembler PGM=ASMA90 instead of

ASMIMSLG its predecessor Assembler PGM=IEV90, a syntax error that

Nov 1, 1996 was overlooked by IEV90 was encountered. The statement

STATS=SKIP following WRITEMSG IMSM101I needs a comma

(STATS=SKIP,) for the continuation.

Thanks to ???, BBK, SPAIN.

Change 14.261A The statement MONICASP=MONACISP; must be changed to

TYPEMON8 MONICASP=MONACASP; to pick up the correct

Nov 1, 1996 temporary value.

Thanks to Alfred Brunner, COMLAB Gmbh, GERMANY.

Change 14.260 DFSMS/RMM SMF records caused INPUT STATEMENT EXCEEDED.

VMACEDGS In the EDGATYPE='O' code, the test IF MOCFLG EQ '08'X ...

Nov 1, 1996 must be replaced with IF MORTYPE EQ '000000000000'X ...

and IF MOVOLNO GT 0 AND MOCFLG NE '08'X THEN DO ...

must be replaced with only IF MOVOLNO GT 0 THEN DO ...

In the EDGATYPE='P' logic, the statement

IF MOVOLNO GT 0 THEN DO _I_=1 TO MOVOLNO; must be

IF MPVOLNO GT 0 THEN DO _I_=1 TO MPVOLNO; must be

Thanks to Siegfried Trantes, IDG Informationsvararbeitung, GERMANY.
Change 14.259 Support for Omegamon for VTAM V200 was not complete; an

VMACOMVT extra four bytes added to IRNUM=13 segment was overlooked

Oct 30, 1996 in Change 14.268. After input of OM13GT, insert:

OM13DT &PIB.4. /*NCP*INTERVAL*DELTA*TIME*/

and add OM13DT to KEEP= list.

Thanks to Neil McMenemy, Royal Bank of Scotland, SCOTLAND.

Change 14.258 Goal Mode variable RPTCLASS was added to PDB.JOBS. The

IMACPDB variable name was added to the _PDB30_4 macro, and to

Oct 30, 1996 the ID statement in macro _PDBSUMS (just like SRVCLASS).

Thanks to Steve Colio, CIGNA, USA.

Change 14.257 The 22 new Shared Paging variables in TYPE71 starting

VMAC71 with SHPBINAU and thru SHPGLOAV need to be INPUT as

Oct 25, 1996 &RB.4. instead of &PIB.4.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.

Change 14.256 Support for APAR OW22209 which (compatibly) adds fourbyte

VMAC21 counters for bytes read and bytes written, replacing the

Oct 24, 1996 two byte counters that could overflow! Also OW25262.
Change 14.255 Formats MGBYTES and MGBYTRT now format Petabytes, because

FORMATS I finally found (in an IEEE Computer Society article)

Oct 24, 1996 that 1000 Tera is 1 Peta!
Change 14.254 TYPE72GO variables R723CSRS, R723CSPA, R723CSPE are still

VMAC7072 wrong! Their informats are &RB.8.6, &RB.8., and &RB.8.

Oct 24, 1996 respectively.

Thanks to Bruce Widlund, Merrill Consultants, USA.

Change 14.253 The JCL example showed //SYSID instead of //SYSIN.

ANALDB2C While minor, its still worth fixing, since a non-expert

Oct 23, 1996 with JCL might not recognize the misspelling.

Thanks to David Middlebrook, National Exchange of Police, AUSTRALIA.

Change 14.252 An invalid RACF segment for RACFTYPE=03 caused STOPOVER.

VMAC80A Although IBM documents a single length of 1 byte, the

Oct 23, 1996 segment had 151 bytes! To protect, insert:

SKIP=RACFDLN-1;

IF SKIP GT 0 THEN INPUT +SKIP @;

just before the END; that ends the WHEN (3) DO; clause.

Thanks to Silva Viviani. Fondo Commune D.C. Rurali Trentine, ITALY.
Change 14.251 INVALID SECOND ARGUMENT TO FUNCTION SUBSTR in DB2 102 due

VMAC102 to incorrect length statement for variable COLLECT, which

Oct 23, 1996 must be $28 instead of $26.

Thanks to Dr. Alexander Raeder, Karstadt, AG, Germany

Thanks to Harmuth Beckmann, Karstadt, AG, Germany
Change 14.250 Cosmetic. TMON for CICS variables CPUTCBTM, UNKNCPTM and

TYPETMON CPUTM were not formatted as TIME12.2, but now are.

Oct 19, 1996

Thanks to Chris Weston, SAS Institute Cary, USA.

Change 14.249 Support for HP's Measureware for AIX has been written and

IMACMWAI tested. The same datasets that were created by HP's PCS

TYPEMWAI for AIX are created, although some variables were deleted

VMACMWAI and some new variables are created.

Oct 17, 1996

Thanks to Lorenzo Wright, NCCI, USA.

Change 14.248 Support for Applied Software's SUPER IND$FILE product's

VMACSUIN SMF record, which creates dataset SUPERIND with stats on

Oct 17, 1996 file trasnfer with that utility program under TSO.

Thanks to George Briscoe, MBNA, USA.

Change 14.247 Support for Oracle 7.0.16 and 7.2.3 is in MXG 14.02 and

VMACORAC later, as there was no change in the format of the SMF

Oct 17, 1996 record created by those releases.

Thanks to Paul Hassett, U.S. Department of Transporation, USA.

Change 14.246 AS/400 variable GDES4 (storage size) was increased by IBM

VMACQAPM causing truncation of the size of installed storage. The

Oct 16, 1996 &NUM.7. after GDES4 was changed to &NUM.10. to correct.

Thanks to Ed Shigo, Nabisco, USA.

Change 14.245 Testing with intentionally duplicated SMF data uncovered

BUILDPDB cases where the PROC SORT NODUP did not remove duplicate

BUILDPD3 observations, because the BY list was underspecified.

BUILD002 For TYPE30MU, add PRODOWNR PRODNAME PRODQUAL PRODVERS to

BUILD003 its BY list.

MONTHBLD For TYPE74CA, add DEVN SMFTIME to its BY list.

WEEKBLD For TYPE74CF, add R744FNAM SMFTIME to its BY list.

Oct 16, 1996 For TYPE74ST, add R744SNAM SMFTIME to its BY list.

For TYPE77, add QNAME RNAME SMFTIME to its BY list.

Thanks to Chuck Hopf, MBNA, USA.

Change 14.244 Change 12.179 was not applied to VMACIMSA, used by the

VMACIMSA ASMIMSLG processing, causing SAP variables SAPTIMTR,

Oct 15, 1996 SAPCPUT, and SAPELTI to be incorrect.

Thanks to Markus Joos, SAS Germany, GERMANY.

Change 14.243 Support for RACF 2.1 IRRDBU00 unload utility has now been

VMACRACF verified for most records. The only change was that the

Oct 11, 1996 variable TIMEOUT in RACF0230 should be INPUT as TIME5.,

instead of the previous ZDB5. input.

Thanks to Peter Cannone, The Great Atlantic & Pacific Tea Co., USA.
Change 14.242 A truncated catalog cell=04 record caused INPUT STATEMENT

VMAC6156 EXCEEDED RECORD LENGTH. The cell has VOLHYKLN=2, but

Oct 10, 1996 that is at the end of the record, and there was no high

key following the key length. Protection added so that

MXG checks to see there is still data in the record prior

to INPUT of VOLHIKY.

Thanks to Raff Rushton, Great Western, USA.
Change 14.241 Support for RACFEVNT=27 (GENERAL AUDIT) creates the new

EXTY8027 TYPE8027 dataset. RACFTYPE=36 is now decoded in SETROPS

IMAC80A (RACFEVNT=24) records.

VMAC80A

Oct 10, 1996

Thanks to Robert Miles Standish, Dean Witter Trust Company, USA.

Change 14.240 IPAC SMF records cause INVALID ARGUMENT TO FUNCTION MDY

VMACIPAC (and print hex dump with "RDS " in bytes 15-18) because

Oct 9, 1996 the SDATE and EDATE are julian dates, not MMDDYY as the

vendor documented.

Thanks to Shaheen Pervais, Trans Union, USA.
Change 14.239 The TAILORNG= JCL parameter added by Change 14.140 is bad

MXGSAS and in the sample MXGSAS JCL Procedureit must be:

Oct 7, 1996 TAILORNG='*.NULLPDS,VOL=REF=*.NULLPDS',

(like the LOAD= and SASAUTO= symbolic parameters).

With the delivered TAILORNG=NULLFILE, a JCL error occurs

(because NULLFILE is a sequential file, not a PDS).

Thanks to Mr. Harnischmacher, TEZ Test und Entwicklungszentrum, GER.
Change 14.238 Archaic IDMS 10.2.1 SMF records caused STOPOVER because

VMACIDMS the subtype 2 record is 4 bytes shorter, and the subtype

Oct 1, 1996 6 is one byte shorter than the code expected.

Oct 8, 1996 At the end of the INPUT for PMHRTYPE=2, change the +12 to

Oct 16, 1996 +8, and change SKIP=SKIP-108; to SKIP=SKIP-104;

I suspect a CA PTF changed the record, because I am sure

this code ran against 10.2.1 records in past years, but

the fix protects both 10.2.1 and 12.01 records.

-In the PMHRTYPE=6 code, replace the line with +1

after the line inputting JRLFILE with:

@; IF SMFHVER GE '1201' THEN DO;

INPUT +1 @; SKIP=SKIP-1;

END;

and change SKIP=SKIP-105; to SKIP=SKIP-104;

-In the PMHRTYPE=18 code, change code now reading:

DBKPGGRP &PIB.2.

+2

DBKKYFMT &PIB.4.

DBKLTYPE &PIB.4.

@;

SKIP=SKIP-14;

IF DBKOWNER='80'X THEN DO;

to instead look like this:

DBKPGGRP &PIB.2.

@;

SKIP=SKIP-4;

IF SMFHVER GE '1201' THEN DO;

INPUT +2

@;

SKIP=SKIP-2;

END;

INPUT DBKKYFMT &PIB.4.

DBKLTYPE &PIB.4.

@;

SKIP=SKIP-8;

IF DBKOWNER='80'X THEN DO;

Thanks to Paul Hasset, U.S. Department of Transportation, USA.

Change 14.237 Candle's EPILOG for MVS records with SM180SUB='RCCH' are

VMACEPMV not the expected record and are deleted. The test for

Sep 30, 1996 'RSRC' was expanded to OR SM180SUB='RCCH' THEN DELETE;

Also, the following INPUT statement is now conditionally

executed as IF LENGTH-COL+1 GE 152 THEN INPUT ....

Thanks to Dan Almagro, Automobile Club of Southern California, USA.

Change 14.236 Support for new RMF type 74 subtype 100 record, called

EXTY74LK the IRLM 'long lock' package, added by APAR OW20579 and

IMAC74 others, that measures held time for IRLM Structure locks

VMAC74 in the Coupling Facility. This APAR has been in FIXTEST

Sep 28, 1996 for quite a while but has been distributed to some

customers, and there is an RMF User report that displays

locks held/waited. New dataset TYPE74LK contains one

observation for each IRLM lock structure name. This code

has not been data-tested as yet.
Change 14.235 Support for several new fields (mostly SMS-related, like

VMXGHSM Data, Storage, and Management Classes) added to the HSM

Sep 28, 1996 MCC record. Additionally the SC,DC, and MC class names

Oct 9, 1996 in both MCC and MCD records are now input only if their

length is non-zero; some zero-length records had hex

zeroes instead of blanks for these name fields. Also, the

INFILE statements needed LENGTH=LENGTH COL=COL added.

Thanks to Terry Duchow, U.S. Postal Service, USA.

Change 14.234 PRO/SMS support in Change 14.207 was incorrect; test data

EXPROSRE for subtype 1 has now been verified, but subtype 2 record

IMACPROS documentation disagrees with actual content of test data.

VMACPROS There are now two datasets created:

Sep 28, 1996 Dataset Subtype Description

Oct 16, 1996 PROSMS 1 Message

Nov 1, 1996 PRORECOV 2 Recovery Performed

Both datasets have been tested with actual data.

Thanks to Warren Hayward, TJX Companies, USA.
Change 14.233 TYPE89 usage data variable MULCURD, for Batch Pipes, was

VMAC89 incorrectly input. The input format of MULCURD is set by

Sep 26, 1996 the value of MULTCURT, but MXG had reversed the INPUT for

MULTCURT 2 and 3. The correct statements should be:

IF MULCURT EQ 1 THEN INPUT MULCURD &RB.8.2 @;

ELSE IF MULCURT EQ 2 THEN INPUT MULCURD &PIB.8. @;

ELSE IF MULCURT EQ 3 THEN INPUT MULCURD &RB.8. @;

The choice of format for MULCURD is set by the product's

FUNCTIONDATA request in its macro IFAUSAGE; previous data

records contained MULCURT=1, which correctly read MULCURD

but BatchPipes sets MULCURT=3, exposing the error.

Thanks to Michael Oujesky, MBNA, USA.

Change 14.232 HSM warning "FSRTVOLS CAN CONTAIN ONLY 30 BYTES" is writ

VMACHSM in error. The two lines preceding the PUT of that

Sep 25, 1996 message, now reading IF LENGTH(FSRTVOLS) GT 23 THEN ...

must be changed to: IF LENGTH(FSRTVOLS) GT 30 THEN ...

Thanks to Robert Miles Standish, Dean Witter Trust Company, USA.
Change 14.231 Change 14.167 was implemented and printed incorrectly.

DIFFDB2 The four lines after that change that start with:

Sep 25, 1996 IF QWHSISEQ GE 1 AND PREVSSID .... must be changed to:

IF SEQCHECK GE 1 AND PREVSSID ....

One source of the skipped sequence numbers (QWHSISEQ) is

CA's INSIGHT monitor for DB2. When INSIGHT is started at

DB2 startup, several ISEQs are skipped, and several ISEQs

are skipped at shutdown. CA is investigating.

Thanks to Tom Parker, Hogan Systems, USA.
Change 14.230 Huron dataset HURN47 had zero observations if there were

VMACHURN no external resource segments at the end of the record.

Sep 23, 1996 After the END; statement that ends DO I=1 TO HU47XSNO;

insert:

IF HU47XSNO EQ 0 THEN DO;

%%INCLUDE SOURCLIB(EXHRN47);

END;

Thanks to Mark Totleben, Farm Bureau Insurance, USA.

Change 14.229 Support for Interlink's Harbor 4.1 SMF record creates

EXHARARC nine datasets for that backup product:

EXHARBKP Subtype Dataset Description

EXHARCON 01 HARBBKUP Harbor Backup Request Info

EXHARDIS 02 HARBREST Harbor Restore Request Info

EXHARFIL 03 HARBARCH Harbor Archive Request Info

EXHARMIG 04 HARBRETR Harbor Retrieve Request Info

EXHARREC 05 HARBDIST Harbor Distribution Info

EXHARRES 06 HARBMIGR Harbor Migration Info

EXHARRET 07 HARBRECA Harbor Recall Info

FORMATS 08 HARBCONS Harbor Consolidation Info

IMACHARB 09 HARBFILE Harbor File Pickup st

TYPEHARB The first two subtypes have been validated with test data.

VMACHARB

Sep 22, 1996

Thanks to Paul Hassett, U.S. Department of Transportation, USA.

Change 14.228 Support for EOS (Change 14.143) wasn't correct; the Audit

FORMATS records were INCOMPATIBLY changed, causing MXG to trash

VMACWSF the WSFAUDIT dataset (but there was no execution error,

Sep 22, 1996 and the other datasets were validly created). There were

additional compatible changes made to EOS (aka WSF2):

New values for format MGWSFAC

New Audit variables in dataset WSFAUDIT:

AUDCFO - Form Used to Create

AUDCUSR - Report Owned By

AUDCTIM - Report Capture Datetimestamp.

New event variables in dataset WSFEVTSC:

ACCIPAD - IP address

ACCLUTY2 - LU Type (VTAM or TCP/IP)

ACCRESP5 - Responses LE 30 seconds

ACCRESP6 - Responses LE 2 minutes

ACCRESP7 - Responses LE 5 minutes

ACCRESP8 - Responses GT 5 minutes

Thanks to Judy Arroyo, Summit Bank, USA.

Change 14.227 Variable R742PSTA in dataset TYPE74PA is now decoded to

FORMATS describe path status, based on APAR OW22735, with new MXG

VMAC74 format $MG074ST.

Sep 20, 1996

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
Change 14.226 DB2 Group Buffer Pool dataset DB2GBPST repeats the first

VMACDB2 buffer pool's data because the pointer was not moved.

Sep 18, 1996 Insert @OFFQBGL @; before the DO _I_=1 TO NRQBGL;

statement, and then remove the @OFFQBGL from the INPUT

statement after that DO statement.

Thanks to Bertolotti Emmanno, Banca Commerciale Italiano, ITALY

Thanks to Burani Daniela, Banca Commerciale Italiano, ITALY
Change 14.225 CICS/ESA 3.3, with optional ONC RPC feature creates a 5th

VMAC110 TCB in SMF 110 Statistics STID=57 records that was not

Sep 18, 1996 expected in MXG, causing MXG to print "UNEXPECTED (NEW)

Oct 9, 1996 DATA WAS FOUND" message on the SAS log. The APAR that

made the change was PN63215, closed in 1994, so it seems

that few sites have the ONC RPC feature on CICS 3.3.

MXG supports the 5th TCB for CICS 4.1 (and the 6th for

CICS 5.1), so only the 3.3 logic had to be corrected.

Oct 9: DSGNTCB is supposed to be the number of TCBs, but

in some CICS 3.3 records, it is zero, so the logic was

revised.

Thanks to Siegfried Trantes, IDG, GERMANY.

Change 14.224 Labels for CONTROL-T variables DENSITY, RECFM, TRTCH, and

VMACCTLT VDENSITY were corrected.

Sep 16, 1996

Thanks to Paul Hassett, U.S. Department of Transportation, USA.

Change 14.223 The INFILE statements for TANDCTLR and TANDLINE must be

VMACTAND LRECL=86 and LRECL=128 respectively. While comments had

Sep 12, 1996 correct values, both INFILE statements had LRECL=342.

Thanks to Alan Phelan, Allied Irish Bank, IRELAND.

Change 14.222 ASM fails; last minute changes were not tested. Find

ASMDALO Find 'R12,R11' and remove the blank after MXGDALO

Sep 12, 1996 Exclude ALL, Find 71,71 'X' ALL and in those four lines

move the X in column 71 to column 72 (so the Assembler

will recognize them as continuations).

Thanks to Chuck Hopf, MBNA, USA.

====Changes thru 14.221 were included in MXG 14.07 dated Sep 11, 1996===
Change 14.221 Support for IMS 5.1 Fast Path records using TYPEIMS7 to

VMACIMS read IMSLOG was added. The ASMIMSLG processing member

Sep 10, 1996 VMACIMSA already supported the Fast Path changes; this is

the same code, moved into the VMAC used by TYPEIMS7.

Thanks to MH, Allied Irish Bank, IRELAND.
Change 14.220 CICS Exclude logic for CICS 4.1 had a missing semicolon

IMACEXCL (the line BASETIME=...) and the IF SMFPSRVR GE 41.1 THEN

Sep 10, 1996 must be IF 33.0 LT SMFPSRVR LE 41.0 THEN ...

Thanks to Pat Quinnette, Principal Mutual Life Insurance, USA.

Change 14.218 Support for SNMP IFENTRY tables from TRENDSNMP product

EXTRSIFE from Desktalk, which polls SNMP agents and builds SYBASE

IMACTRSN tables, either under UNIX or WINDOWS-NT. MXG can execute

TYPETRSN either on an ASCII platform that connects to the SYBASE

VMACTRSN system and uses PROC ACCESS to directly convert the data

Sep 10, 1996 into SAS datasets, or instead, you can use the SYBASE BCP

utility to create an ASCII flat file that MXG can read on

the PC/Workstation, or that ASCII file can be shipped to

MVS and processed there. Macro _SIFEN uses PROC ACCESS,

Macro _TIFEN uses INFILE/INPUT to read BCP data. This is

the preliminary implementation and has been tested, and

will read either the Rate table or the History Table.

This code does no date/time selection; you get all of the

data in the SYBASE table, so you will need to select the

desired data. Future enhancements to provide selection

of "just yesterday's" data are planned, as are support

for additional SNMP tables and RMON data. Benchmarks on

my 64MB 166MHz Pentium with Windows NT 3.51 reading 20MB

SYBASE table that contained 6 days hourly data from 1000