Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document Parallel Test Approach I. Introduction The purpose of this document is to explain the approach that will be used for a Parallel Test phase of the University of Michigan HE Upgrade Project. This document outlines the following: Parallel Test objectives and scope Parallel Test comparison tools Parallel Test entry and exit criteria Parallel Test calendar Parallel Test team Parallel Test issues and risks Parallel Test issue resolution process Parallel Test assumptions II. Parallel Test Approach A. Objectives The primary objective of Parallel Test is to identify discrepancies in the Gross-to-Net calculation between versions of Michigan’s PeopleSoft system. All data will have been entered using the PeopleSoft system and will have been converted into the Parallel environment via the conversion script to be used for production implementation (no conversion from legacy data is required). Table 1 includes the components of the gross to net calculation that may cause discrepancies between paychecks created in the v7 PeopleSoft Payroll System and paychecks created in the v8 PeopleSoft Payroll System. The parallel testing process tests these components and resolves any discrepancies between the two. Resolutions may come in the form of: Updating software configuration Correction of work units (e.g. fixing custom programs that impact gross-to-net) PeopleSoft correcting delivered source code Correcting data conversion scripts Explaining and accepting the discrepancy. Document1 Page 1 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document Other normal biweekly and monthly processes will be run and results checked, but will not be compared against production data. E.g. Paycheck distribution, gross pay register, G/L interface, leave accrual, etc. Table 1 Gross Earning Calculations – verifying the employee’s gross earnings tests the following: delivered/upgraded gross earning calculation routines correctly calculate an employee’s gross pay modified/custom gross earning calculation routines correctly calculate an employee’s gross pay earnings table is set up correctly Deduction and Garnishment Calculations – verifying the employee’s deductions and garnishments tests the following: delivered/upgraded deduction and garnishment calculation routines correctly calculate all types of deductions (e.g., before tax, after tax, employer paid, etc) modified/custom deduction and garnishment calculation routines correctly calculate all types of deductions deduction table is set up correctly garnishment rule and disposable earning definition tables are set up correctly benefit plans and benefit programs are set up correctly Tax Calculations – verifying the employee’s taxes tests the following: delivered/upgraded tax calculation routines correctly calculate federal, state, and local taxes modified/custom tax calculation routines correctly calculate federal, state, and local taxes federal, state, and local tax tables are setup correctly Time & Labor – verifying the loaded time data tests the following delivered/upgraded Time Admin process correctly loads time into the system modified/custom Time Load Interface correctly loads time into the system B. Scope In order to fully test Michigan’s Gross-Net-Calculation for both of its pay frequencies, five separate parallel tests will be executed. Parallel testing uses historical payroll data from the current version of the PeopleSoft Payroll System. In other words, prior to the start of parallel test, the pay periods have already been processed in the Payroll System. First, this allows the parallel test team to extract the historical payroll data and prepare it for comparison to PeopleSoft data without interfering with the current production system’s operations. Second, flexibility is built into the parallel test schedule because it is not dependent on the production system’s payroll processing schedule. Document1 Page 2 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document The parallel test will not test each employee in the system, since doing so would require replication of manual keypunching done by timekeepers. Instead, a cross-section of employees has been chosen, which will include most of the University’s employees. The employees in parallel test will meet one of the following conditions: 1) Loads to Time and Labor via interface file (including Corpay – scanned timesheets, as well as interfacing departments). 2) Does not exist in Time and Labor. 3) Is in workgroup PAEXHOSP. This workgroup includes many MPP employees, most of which are not eligible for TRCs such as Overtime or Shift, which would affect their Net Pay. Even though their time will not be loaded into Time and Labor, this should not affect their pay. The Parallel Test will load Convenience Deduction data. Cycle 4 will include manual changes made to paysheets, since Payroll Office staff will perform dual entry for these changes in the Cycle 4 parallel tests. BWR employees will only be tested in Cycle 1. Prior Period adjustments will not be tested in Parallel test; they will be included in System Test. System Test will also test keypunch data entry through the payroll cycle. HE 8.0 Upgrade Parallel Test Scope Cycle 1 Time files Document1 Time Load Interface Time Admin appengine Paysheet Paycalc G/L Interface Sponsored Research Page 3 of 13 Cycles 2, 3, 4 Pay Confirm Last Updated - 2/9/16 Print Paychecks Direct Deposit Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document C. Parallel Test Cycles: The following test cycles have been defined for Parallel Testing o Cycle 0 – Mock Parallel Verify that PARA(for new PS version) has been setup correctly, and that programs/tables/data have been migrated completely. Data will not be reconciled at this point. o o o o Document1 Should be completed prior to “real” start of Parallel Test. Cycle 1 – Biweekly without Time Load XX/XX/XX pay period end date Based on date of production cut of data, Time WILL already have been loaded into payroll; so custom Time interface WILL NOT need to be run. Verify that Payroll-only functionality has not been adversely affected by upgrade. Cycle 2 – Biweekly with Time Load XX/XX/XX pay period end date Based on date of production cut of data, Time will NOT have been loaded into payroll; so custom Time interface WILL be run. Verify that T&L functionality, in addition to payroll, has not been adversely affected by upgrade. Cycle 3 – Monthly with Time Load XX/XX/XX pay period end date Based on date of production cut of data, Time will NOT have been loaded into payroll; so custom Time interface WILL be run. Includes G/L interface, and sponsored research. Verify that the upgrade did not adversely affect Monthly pay processing. Cycle 4 – AFTER NEW CUT OF PRODUCTION DATA (from XX/XX/XX to XX/XX/XX) AVAILABLE Re-run Cycle 2 (Biweekly with Time Load) for XX/XX/XX pay end date for Biweekly Re-run Cycle 3 (Monthly with Time Load) for XX/XX/XX pay end date for Monthly Central Payroll and Hospital Payroll offices to assist in reviewing and verifying results. Page 4 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document D. Parallel Test Comparison Tools The Parallel Test Comparison Toolset is a set of PeopleSoft queries, on-line objects, and SQR’s that compare the Gross-to-Net calculation between v7 PeopleSoft Payroll system and v8 PeopleSoft Payroll system. These queries, on-line objects, and SQR’s will be created during the Parallel Test Planning and Prep phases to meet Michigan’s requirements. A DB Link between the Production environment and Parallel environment will be created for this purpose. There following Parallel Test Comparison Tools created for this effort: Check Payable Time Query Gross to Net Comparison Report with Earnings/Deductions/Taxes Detail Check Payable Time Query This query will validate the data in TL_Payable_Time (V8) against the Tl_Empl_Dlyerns and Tl_Empl_Dlysum (V7) after the Time Administration and Load Time and Labor is run. This Query will print only the discrepancies between V8 and V76 data into an Excel Spread Sheet. This will help isolate data errors resulting from the Time & Labor process, which can then be corrected prior to the start of payroll processing. Gross to Net Comparison Report This SQR will create a detailed comparison report of the differences between v8 and v7. The program will first compare the net pay amounts in each system. If a discrepancy exists, total earnings, total deductions and total taxes will then be compared. If either of those components does not match in both environments, a detailed listing of records will be written from each system. The resulting report will be an Excel spreadsheet to allow for quick and easy sorting and searching and resolving discrepancies. The following fields (order TBD) will be included: Employee ID Paygroup Department Code Pay Period Begin and End Dates Total Earnings, Deductions, Taxes, Gross Pay, Net Pay from v8 Total Earnings, Deductions, Taxes, Gross Pay, Net Pay from v7 Total Earnings, Deductions, Taxes, Gross Pay, Net Pay Differences Earnings Detail The following fields (order TBD) will be included: Employee ID Paygroup Department Code Pay Period Begin and End Dates Document1 Page 5 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document Earning Codes and Amounts Differences between v8 and v7 Deduction Detail The following fields (order TBD) will be included: Employee ID Paygroup Department Code Pay Period Begin and End Dates Deduction Codes and Amount Differences between v8 and v7 Tax Detail The following fields (order TBD) will be included: Employee ID Department Code Pay Period Begin and End Dates Tax Codes and Amount Differences between v8 and v7 Document1 Page 6 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document E. Entry and Exit Criteria Parallel test entry criteria define the tasks that must be satisfied before parallel testing can be efficiently executed. Parallel test exit criteria define the successful conclusion of the parallel testing stage. Entry Criteria If one or many of these tasks are not complete, parallel test can not begin and the start date will be delayed which may delay the parallel test completion date and ultimately the overall conversion date. Table 2 defines the parallel test entry criteria. Table 2 1) The Parallel Test environment is established for sole purpose of parallel test and includes all custom objects. 2) The control tables are loaded into the parallel test environment. 3) Potential “CPU-hog” programs have been tuned to run as efficiently as possible: Interface Load Time, Time Admin, etc. 4) The parallel test team members have been identified and have dedicated time to work on the parallel test effort. 5) The parallel test team has access to the TI Resolution team. Discrepancies will arise that require PeopleSoft COBOL, SQR, and Online knowledge and fixes need to be completed in a timely fashion to keep parallel test on schedule. Exit Criteria Parallel test is not complete until all of the exit criteria have been satisfied. If any of the parallel test exit criteria are not satisfied, unexplained discrepancies may exist in the parallel test employee population’s gross to net calculation. As a result, the parallel test team will not have the data that it needs to contribute to the “go live” decision. Table 3 defines the parallel test exit criteria. Table 3 1) All discrepancies have been explained or corrected, and documented. 2) All earnings calculation discrepancies have been explained or corrected, and documented: 3) All deduction and garnishment calculation discrepancies have been explained or corrected and documented. 4) All tax calculation discrepancies have been explained or corrected and documented. Document1 Page 7 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document F. Parallel Test Calendar The parallel test calendar shows the cycles and schedule (HE 8.0 Upgrade Parallel Test Schedule.xls) 5/26 6/2 6/9 6/16 6/23 6/30 7/7 7/14 7/21 7/28 8/4 8/11 8/18 8/25 9/1 9/8 9/15 9/22 9/29 10/6 10/13 10/20 10/27 11/3 11/10 11/17 MOCK PARALLEL - 3 wks biweekly, monthly, time load REFRESH ENVIRONMENT CYCLE 1 - 3 wks 12/27/02 biweekly, no time load CYCLE 2 - 3 wks 1/10/03 biweekly + time load Apply Tax Update 02-G CYCLE 3 - 3 wks 1/31/03 monthly + time load, G/L, sponsored research TEST MOVE #3, new data from 7/31/03 for Parallel Test Parallel Resources move to Systest Apply Tax Updates thru 7/31 Prod CYCLE 4.2 - 3 wks re-run cycle 2 for 8/8/03 biweekly CYCLE 4.3 - 3 wks re-run cycle 3 for 8/31/03 monthly Apply remaining Tax Updates Document1 Page 8 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document G. Team Structure The following chart lists the team members who will be involved in the Parallel Test. The product managers of each area (1st row listed) will have the final say in how a discrepancy is to be resolved or explained, with Scot having ultimate signoff. We would like to have Central Payroll and Hospital Payroll offices involved in reviewing and verifying results during the Parallel Test phase. Our goal is involve these groups during Cycle 4 at the latest. Payroll/Time & Labor: Scot Yoas, Carrie Shumaker, Ruby Nugent, Arun Madadi, Nancy Goff, Kristin Lorenz, Dave Jamison, TBD Parallel Consultant; all PYTL team members will be on the TI Resolution team. Benefits: Denise Stegall, Amy Lawrence Human Resources: Judy Aldrich, Val VanHaaften Campus Community: Karen Kuffner, backup Higher Education: as needed Payroll Office: as needed Hospital Payroll: as needed Document1 Page 9 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document III. Status and Reporting Metrics A. Defect Tracking Approach During all phases of system testing, defects will be logged into the Lotus Notes Testing Incidents database, including performance tuning needs. The database will provide the necessary information for developers, as well as provide management with reports. The developers need the name of the reviewer, date discovered, the requirements, test case id, severity code, and any other relevant information. All high and medium priority issues must be resolved or have a work-around in place to go-live. To ensure timely resolution of cases submitted to PeopleSoft during the system test phase, the SIT lead will be provided a contact at PeopleSoft. B. System Test Phase Tracking Approach – Metrics To Utilize System test will be tracked by phase and within each phase the business processes by module. Leads will be responsible for ensuring that time is entered into Planview and all stats are reported to the system test phase lead in a timely manner. The format of the status reports being utilized for design and development will be used for system testing, as well as the reusable test plan reports. There may be additional reporting requirements, as well as, the CPUs may have additional detailed reports. Metrics that will be tracked include but are not limited to: 1. Number of conditions, by priority (ABC), that are in progress, in error, passed, deferred, or not started. 2. Percentage of conditions tested. 3. Percentage of conditions passed. 4. Percentage of A, B, and C conditions. (Including % of each priority that has been tested and passed.) 5. What cycle/sub-cycle the conditions are in. 6. Total number of conditions per cycle/sub-cycle. 7. Total number of conditions per priority. 8. Percentage of Effort Complete 9. Percentage of Effort Earned. 10. Percentage of Effort Remaining. 11. Number of Open Testing Incidents – by priority. Document1 Page 10 of 13 Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document C. System Test Status/Issues Meeting During system testing, it is recommended that a weekly status meeting be held at the same scheduled time on the same day each week. This meeting will support a communication process to inform the project team on the testing status, problems or issues, and changes. The Agenda will include a review of open high and medium issues, the overall system test plan, and any action items from the previous week. Meeting minutes should be taken that include a list of attendees, discussion points, and action items. Developers and reviewers should be represented in the meeting, as needed. IV. Assumptions, Issues, and Risks for Parallel Test Approach A. Assumptions The following assumptions have been made regarding this Parallel Test: 1. The Time and Labor portion of the test will begin with the Interface Time Load, not with files being sent from vendors. 2. Manual entry will not be done. Only the employees coming through the time files will be parallel tested. 3. Parallel Test needs to be supported like a production system – with quick turnaround time from all supporting teams (functional and technical). Turnaround time for all requests must be less than 24 hours. 4. The Parallel Test team will approve all migrations into the Parallel environment prior to the move. This is needed so that migrations do not interrupt a paycalc (or other process) that is in progress. This has been raised to the STAT project team to include a workflow specifically for Parallel Test. 5. Cut of Parallel Test data already has 12/27/02 pay end date time loaded. B. Issues and Risks The table below lists the issues and risks associated with the University of Michigan Parallel Test. Accompanying each issue is an associated risk if an individual issue is not resolved. Issue Document1 Page 11 of 13 Risk Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document Issue Risk Parallel Test and System Test 1. schedules completely overlap (both are scheduled to start in June), with Mock Parallel preceding System Test by a few weeks. Parallel test is executed in its 2. own environment. If this environment is not set up or configured correctly, parallel test can not start on time. The control tables need to be 3. loaded into the parallel test environment. Parallel test team members with 4. the prerequisite skills and knowledge need to be identified and be dedicated to the parallel test effort. The parallel test team needs 5. access to the Fix-It team. Discrepancies will arise that require PeopleSoft knowledge and fixes need to be completed in a timely fashion to keep parallel test on schedule. The parallel test team may spend time investigating discrepancies that result from problems that should have been identified and fixed during system test. Document1 Page 12 of 13 Any other use of the environment being used for parallel test will impede the progress of parallel test. Parallel test can not begin until the payroll and benefits control tables are loaded into the parallel test environment. If the parallel test team is not a full time dedicated team, the parallel test may not be complete before the decision needs to be made as to the status of the “go live” date. If the TIs that result from parallel test are not fixed in a timely manner, the parallel test may not be complete before the decision needs to be made as to the status of the “go live” date. Last Updated - 2/9/16 Information and Technology Services PeopleSoft Upgrade Project M 5000A2 – Parallel Test Approach Document C. Issue Resolution Processes In order for the Parallel Test execution to run as smoothly as possible, and the resources to be used as effectively as possible, the following processes should be followed when issues arise: Functional Issue Escalation When issues that cross functional teams arise, notify Scot, and then call primary, if no answer, call secondary. Follow-up with email to both, .cc Scot. If data changes need to be made to any environment, a TI should be created for that request. Technical Issues When technical issues arise (e.g. programs not running, version control, etc.) the TO group will be notified via standard procedures. Email during business hours, call 3-4000 during off-hours. The only difference is that the PARA8 environment will be supported at a higher-level than standard testing environments (see Assumptions). Migration Requests Migration requests will follow the standard OMR/STAT procedure. The only difference being that a Parallel Test team member will approve all migrations into the environment prior to the actual move. (See Assumptions) V. Deliverables Work Breakdown Structure (WBS) - updated business processes and adjusted hours. This is required at the beginning of the system test phase, in order to put the WBS into planview for tracking and reporting. M571 – System Test Plan o Master reusable test plan (RTPs) with appropriate rating scale assigned to each business process. o Test scripts and process flows as necessary o All internal and external touch points/interfaces and key contacts must be identified. STAT - OMR TI (Testing Instance) Database (Resolution of issues or work around in place) M539 – Parallel Test Status Report Document1 Page 13 of 13 Last Updated - 2/9/16