Re: How about employing TLS for private DBus connections ?
rony <[email protected]>
| Newsgroups | gmane.comp.freedesktop.dbus |
|---|---|
| Organization | WU Wien |
| Message-ID | <[email protected]> |
Simon: On 08.10.2018 12:15, Simon McVittie wrote: > On Sat, 06 Oct 2018 at 15:27:42 -0700, L A Walsh wrote: >> On 8/24/2018 3:43 AM, rony wrote: >>> TLS would take care of encryption already. >> --- >> While TLS may be needed over some networks, there are private or >> home networks, where it wouldn't be needed. >> >> Why can't the networks be graded as private, corporate or public >> so rules can be applied as needed for the specific security situation? >> >> This has worked for years on MS networks, so certainly it >> should be good enough for linux. > This is not really comparable, because Microsoft control every aspect > of their operating system. The maintainers of dbus don't control every > aspect of a single Linux distribution, let alone every implementation > of Linux, *BSD, Darwin and all the other platforms the reference > implementation of D-Bus is meant to work on. > > There are certainly implementations of networks marked as private, > corporate or public on Linux systems - firewalld does that, for example - > but not every user of dbus uses firewalld (and firewalld itself relies > on D-Bus, so it would be a circular dependency if dbus relied on it). > > Microsoft presumably also pay multiple people to develop and maintain > their IPC systems; I am not aware of any dbus contributor who works on > it full-time. (I certainly don't.) > > If this is not acceptable to you, either don't use dbus, or help us. > When we try to set a realistic scope for what we can support, telling us > that we are not doing enough is not constructive. > > smcv that is very understandable! If it is o.k. with you and others in the D-Bus community I would try to reserve enough time over the Christmas vacation to attempt to add (experimental) TLS transport for evaluation. The idea being to allow an address "tls:host=...,port=...,keystore=..." instead of "tcp:host=...,port=...". To follow the KISS principle the creation and deployment of the keystore/certificate that is needed on the server and all client sides, would not be part of D-Bus, but part and responsibility of the deployer. The result should be that no one but those who have access to the same keystore/certificate would be able to communicate with each other (shared secret). As long as that keystore/certificate is kept at a safe place on each machine, no one should be able to break in. Having the ability to deploy private D-Bus client/server applications I really think this would be a real great boon for application developers (and for D-Bus as well), if they became able to take advantage of the D-Bus infrastructure to help them create simple client-server applications and be able to mix and match Linux with Windows and MacOSX (addressing mostly small to medium businesses). There are two specific problems that I might face: - enough time to meet that challenge (next possibilities would be February, Easter vacation and summer), - enough specific knowlege: not being a system programmer, let alone a Linux system programmer, I would benefit a lot, if you or other D-Bus developers could point me to the code that resolves the tcp address and sets everything up for a private D-Bus connection (and warnings, if you can foresee/speculate/know of potential problems). Would that be something that would be interesting, helpful to the D-Bus community? ---rony _______________________________________________ dbus mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dbus