Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

IPv4/IPv6 handling #17

Open
reisub-de opened this issue Sep 22, 2017 · 6 comments
Open

IPv4/IPv6 handling #17

reisub-de opened this issue Sep 22, 2017 · 6 comments
Labels

Comments

@reisub-de
Copy link
Contributor

We need to clarify IP version handling.

  • Do we set global requirements? (If you want to use IPv4, every node has to speak IPv4 and vice versa)
  • Do we support heterogeneous networks? Then we eventually need to split the paths into IPv4 and IPv6 compatible.
  • Do we allow to advertise IPv4 networks but the routers must be able to support IPv6, too? So we can find paths for tunneling and simultaneously allow for advertising these subnets over DMPR. I think this by far exceeds the scope of DMPR.

Global requirements are the easiest to implement and specify but limit the usefulness of DMPR while in heterogeneous networks we would need to advertise one path for IPv4 and one for IPv6

@hgn
Copy link
Member

hgn commented Sep 22, 2017

This heterogeneous setup is currently supported. You can transport v4 and v6 prefixes. Fine.

Then for the transport, if we have v4 we mcast on v4 address. If we have v6 additional enabled we additionally mcast the data on a v6 mcast socket. This doubles the data but should work

@reisub-de
Copy link
Contributor Author

The problem is that a path via an IPv4-only Router is not valid for an IPv6 packet

@hgn
Copy link
Member

hgn commented Sep 22, 2017

Yes, that is a subsequent challenged, solved subsequently e.g. setup ipv4 tunnel.

@reisub-de
Copy link
Contributor Author

The tunnel would be below and therefore transparent to dmpr, wouldn't it? Then we still have the problem of paths between IPv4-only and IPv6 routers

@reisub-de
Copy link
Contributor Author

ipv46

A advertises a route to D.
To A comes an IPv6 packet for D, it gets forwarded to B and then dropped because B has no IPv6 hop to C

@hgn
Copy link
Member

hgn commented Oct 22, 2017

I think we should pause this for a while. To solve this issue correctly - and never address this again - we should:

  • read other routing protocol transition strategies (BGP, OSPF, ...)
  • Keep the transition logic/functionality as small as possible. E.g. if a tunnel is required to fulfill v4/v6 transition there is no need to keep this into the protocol
  • we should speak with routing && transition experts how to cope with this.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants