Attention is currently required from: lynxis lazus, pespin.

laforge has posted comments on this change by pespin. ( https://gerrit.osmocom.org/c/osmo-ggsn/+/38500?usp=email )

Change subject: doc: Document MTU features in User Manual and example config files ......................................................................

Patch Set 6:

(2 comments)

File doc/manuals/chapters/running.adoc:

https://gerrit.osmocom.org/c/osmo-ggsn/+/38500/comment/a1314577_90067642?usp... : PS5, Line 57: When running osmo-ggsn, the user must take network Maximum Transmission Unit : (MTU) into consideration, and configure it based on network specific setup.

Indeed there's nothing wrong with IP fragmentation. The problem is, as usual, broken middleboxes. […]

I'm not aware of any common breakage preventing transporting fragmented IP packets across networks. The only frequent problems I know about is PMTU discovery failures, which really are just ICMP filtering.

https://gerrit.osmocom.org/c/osmo-ggsn/+/38500/comment/797b2ff2_0ad50b0c?usp... : PS5, Line 122: OsmOGGSN

Sure, but it is still interesting to document imho since: […]

"95% of the traffic will be TCP": I think that's no longer the case today. There's bittorrent, openvpn, ipsec/esp, wiregurad, RTP (for webrtc, whatever), there's QUIC, etc. - if anything I think the percentage of TCP traffic of the users on the public net is likely decreasing than increasing. I have not investigated how well those various other protocols handle MTU issues.

"increasing the outer MTU" that's basically what 29.060 annex C states: (paraphrasing from memory): "When you cannot enlarge the outer MTU across the entire transport network as needed, then you should consider reducing the inner MTU to avoid fragmentation overhead"