Security Considerations with Object Connect

Hello,

I did some research on the subject but resources are pretty scarce. This is kind of a follow-up to my question about APPN/APPC security a while ago. Again, my findings are valid for V4R5 at the moment. I did not yet do verification on newer releases. The authoritative source about the topic can be found here: https://www.ibm.com/support/knowledgecenter/ssw_ibm_i_72/rzarm/rzarmobjcht.htm

IBM seems to assume that if you are permitted to start SAVRST* on a source machine A, a corresponding job will be spawned on machine B. QSECOFR is allowed to do this by permissions of the SAVRST* commands.

Apparently, if two machines are networked via APPC, there's no means against malicious overwriting of data from a peer machine. Denying the automatic creation of APPN *CTLD in the corresponding *LIND helps only against new peers. If I *want* to have an APPN connection to a peer machine, it's apparently all-or-nothing. I can't see some obvious way to restrict incoming communication requests from Object Connect Services.

Lately, I learned from a comment in here about exit points. And I know about job routing. Does anyone have an idea how to use these facilities to restrict Object Connect communication requests? Is it possible to hook up an external program, written by me, to an exit point, verifying the location name of a request (and maybe some more parameters) and permit or deny access (as it's possible with QIBM_QTMF_SVR_LOGON for FTP)? I couldn't find something APPC related in WRKREGINF, though.

Ideas in which direction to search welcome!

Please refrain from suggesting "Don't use SNA". There are services which are either easier to handle, or only available with SNA. Thanks.

:wq! PoC

PGP-Key: DDD3 4ABF 6413 38DE - https://www.pocnet.net/poc-key.asc