SAP TDMS vs Accenture Clone and Test

[Former Member]

For our SAP HR implementation, we are using Accenture Clone and Test tool for our test data preparation. We are looking at whether TDMS is able to complement this in the refresh of the test system from Production.

What advantages does TDMS have over Accenture Clone and Test?

[gerard_white]

Hi Gordon,

since I'm answering from SAP I will just mention some of the strengths of the TDMs HCM offering. Perhaps there's another customer with experience of both Accenture's Clone and Test and SAP's TDMS HCM who can help you more directly with your comparison question. We see some of our strengths in the following:

TDMS HCM is completely integrated in the SAP TDMS basis technology allowing transfer of high data volumes. In a recent customer run, for example, 14,000 PERNRs along with PD, PA, payroll and time data were transferred in one run. Extensive scrambling was active during this. No portioning of this data was necessary.

TDMS HCM is an SAP standard product and therefore fully supported by SAP. The customer feedback to this support has been positive.

TDMS HCM offers simple process trees for each activity step and also sophisticated status management capabilities. Each activity can be thus protected by authorisations and is extensively logged.

TDMS HCM offers a specific role and autorisation concept both at TDMS activity level and at SAP HCM level.

TDMS HCM is a global solution providing country-specific content for data selection and data scrambling.

Please don't hesitate to contact us directly if you need more specific information. Perhaps other forum users could help you with their impressions and experiences.

Gerard White.

[Former Member]

Gerard,

We are currently in the requirement gathering but have not yet analysed all the tools available in the market. TDMS is a desired solution since we have SAP HR. Can you please share any documents that might help us in our analysis of TDMS. Also, the biggest concern at this point is the traceablity w.r.t the downstream systems post a system-refresh which is expected every six months tops. How does TDMS handle this issue?

Appreciate your time on this! Thanks!

[gerard_white]

Hi,

sorry for the horribly late reply. I'm no longer directly involved in TDMS development (and wasn't in this area for ages). I suggest I get in touch with the new owners of TDMS HCM and ask then to help you further.

Once again apologies for the delay.

Thanks and regards,

Gerard.

[suman_pr]

Hi,

Please find below the answer to your queries:

1. There are several documents available in the Service Market Place regarding the various scenarios supported by TDMS which you may refer to understand the product better. To start off, you may refer to the master guide and the operations guide to get an overview of TDMS.

2. Could you please elaborate on the second question regarding traceablity so that I may provide you with the right answer?

Kind Regards,

Suman

[Former Member]

Hi Gerard,  I have looked through a lot of discussions for TDMS, trying to get idea of how long it would take to copy large amount of data from production to a test client, but haven't seen anyone mentioned the useful performance data, until I saw your post. We use TDMS HCM for copy PD & PA. Most of time the copy data is for production issues and data amount is very small. But we've always been asked what if a large trunk of data is needed, can TDMS do the work within the reasonable time frame. I know you indicated you are no longer in TMDS aera, but could you just tell me how long did it take when you performed 14,000 PERNR copy (PD & PA) in one run? How much data involved here in terms of time, e.g. these PERNRs have data since what year?  Thanks so much, Li

[gerard_white]

Hello Li,

this is really difficult to answer since there are a number of dependencies. For example, is payroll data included? If so, how often does the payroll run? In many countries it's once a month but in others such as the USA it's often more frequent and there are more recalculations. Should time events also be copied? The TEVEN table is potentially massive. Should the data be scrambled? If so, then this adds more time.

As you mentioned, I'm no longer involved with TDMS development and I know that certain performance improvements have been made. I shall ask my TDMS colleagues to try and help with further information.

Regards, Gerard.

[Former Member]

Hi Gerard,

Thank you so much for the quick response, I really really appreciate it! Yes, I totally agree with you that it is very difficult to give this type of estimates. I actually tried copying 150 PERNRs in a single run yesterday, finished in 2.5 hours (the sender is our production system). The step taking most of time was data selection, 1 hour and 34 minutes. The next time consuming step was data deletion in receiver system, 40 minutes.

Here are some details for our copying. Payroll data is included, and for the 150 PERNRs run, I used partial start date 1/1/2012. Our payroll run is weekly (not everyone is in a weekly run though). But we don't have time event (TEVEN is empty). Other data with large number of records are in PA2001, PA2002, ASSHR, ASSOB, COIFT, PAYR, REGUH and tax tables, all of them are included in copying. Scramble is also required, but I noticed all phases related to scramble do not take long time.

I only need a very high level of estimates. For example, if I put 5000 PERNRs in a single run, could it take up to one day to finish? 2 days? The most concerning point is data selection step in sender system since it is our production system. Although it only needs one background job, but does comsume memory and CPU, and certain amount of space for large number of PERNRs.

Thanks again,

Li

[gerard_white]

Hi Li,

2.5 hours for 150 PERNRs sounds rather slow. I asked the TDMS HCM team to try to help you our but they seem not to have had time so far. They may have some better ideas but somethings you may considers are the following:

- authorisations: with the authorisation object TDMS4HCM01 you can control the granularity of the check. The more precise the check is the longer it will take. For this reason, the checks are rather superficial when using the "Expert Mode" program.

- You mention using a payrol cut-off date. You can set technical switches so that other, time related tables with typically large volumes are similarly "cut off". Examples of switches are the "USE_BEGDA_PA2001" or "USE_BÉGDA_PA0015".  I'd have to check but I think these would also reduce the number of ASSHR and ASSOB records too since INFTY is a field in them and these INFTYs are ones influenced by the cut-off date.

Could you use the Expert Mode to do the transfer? Alternatively, do you always need to copy all the OM data? Does the organisation alter so much? Could you split the coping of PA and OM?

Anyhow, I hope the TDMS colleagues can help further. Perhaps they could tell you about a table that can be maintained to alter the select strategy of typically large tables. This could also accelerate matters.

Regards, Gerard.

[Former Member]

Hi Gerard,

Thank you so much for taking extra time out of your regular job to answer my questions. Here are some more details based on the information you listed:

1. Authorizations: I don't know much about this. When checking the user ID dedicated to TDMS copy, I see object TDMS4HCM01 is included in "Control for Granular Authority Check" for both PA and PD (* to *).

2. The payroll cut-off date is automatically applied to all time related infotypes and tables (perhaps it's because we use option "Use current PA selection"?), i.e. records before the cut-off date are not transferred. But we do know when TDMS selects data from ASSHR, it does not take into account of PERNRs we put, instead, it pulls all infotypes first. We had to create a custom index on field INFTY to reduce time spent on this one select from over 4 hours to under 30 min.

3. We've never used "expert mode". Is there any benefit using that?

4. For OM data, we only copy HRP1000 and HRP1001 for CP and S, P-CP, P-CP, P-S. It doesn't have much data at all. Sometimes we even use PA solution which only brings CP and P-CP.

Based on what you have provided, I see there could be area we may improve the performance. If TDMS team can give some suggestion, I'll really be grateful! Should I open OSS or just wait to see when they have time?

Best,

Li

[Former Member]

Hello Li,

I would request you to raise a OSS message for the same. The TDMS team would have to look into the system and see which table consumes the maximum time and how can this be enhanced/Improved.

Regards,

Rishav Surana

[Former Member]

Thanks so much Gerard for your help! We will pursue OSS and get help from TDMS team.

Li