location of repository

hi all,

i already sent this to SAG-L last week, before i realized that there actually is a more appropriate forum… :slight_smile: anyway, here is the question:

i have started a little investigation into the possible use of Natural Engineer, and i have a question about the location of the repository. As i understand it, you can have the repository in one of 3 locations:

  • on the mainframe
  • on the network
  • on each user’s own PC

now each of these solution has it’s own requirements of course, but i’m wondering what the differences are in respect to performance ? Is there a big difference in performance between having it on the mainframe and having it on your own pc for example ?

And are there any important (dis)advantages to these alternatives that i should be aware of ?

Is there somebody who can say some wise words about this ?


Don’t know about wise words but can give some pointers.

First - it really depends on what you are ultimately aiming to achieve. For instance, if there are a team of people using NEE then it would make sense to have 1 central NEE repository. This would then indicate either a windows server or mainframe install.

If you have a central NEE repository, then the next question is how to access it. You could use Entire-Network, but this would be slow as it is returning the Adabas records rather than results. You could use RPC connectivity - depends on your env. You can also use SPoD (NDV) to access the NEE repository (preferred solution and works on Windows Server or Mainframe).

If your source code (FUSER) is on the mainframe, then I would suggest Adabas on Mainframe & utilise SPoD (NDV). Your existing Adabas backup procedures are already defined without creating new ones for open systems (windows server or individual PCs).

Performance considerations come after you establish the best fit for your org, for instance avoid Entire-NetWork. SPoD (NDV) will result in probably the best fit (dependent on project teams using NEE) and reduce maintenance on installation & upgrades in future (updating one server rather than separate PCs).

Going further, perhaps a Terminal Server would be a good solution - one installation!

Let me know if you need more information.
Regards, Brian

thanks for the info Brian…

We do have our sources on the mainframe (z/OS, if that makes any difference) and we have SPoD, so what you say sounds good. I guess teamwork would be the main goal, so a central repository would actually be the best i think.

I’m just a little confused about one thing though: if i read the manuals correctly it says that for NEE on the PC to get direct access to the sources on the mainframe (to load into the repository) you need NEE on the mainframe to be present. Is that correct ? I was just just wondering why that is, since SPoD obviously can get to the sources via the NDV, so why can’t NEE do the same ?

and another thing i noticed: in all the nice PDF’s i’ve seen about the new NEE, it has this beautiful user interface (Silverlight based for an even better user experience, according to the commercial talk) but my version ( still has this ‘boring’ standard windows interface, what’s up with that ?


SPoD (NDV) needs Natural on z/OS to get the source code.
Natural Engineer would need to be installed on z/OS as the NDV provides the communication mechanism between the PC client GUI and the z/OS repository to access the NEE Natural functions that execute on the z/OS server.
HTH, Brian