Jim Thompson <j...@netgate.com> writes:
On May 28, 2019, at 10:19 AM, Christian Hopps <cho...@chopps.org> wrote: Paced tunnel packets on high speed links. So timing considerations -- a 10GHz link can send a 1514 MTU (1548B) Ethernet frame every ~1230ns, a 40GHz link every ~308ns, and a 100GHz link every ~123ns. Probably can't get all the way to the 100GHz performance right now, but I want to get as close as I can. :)I’m struggling to understand how you get from 1514 MTU (isn’t that 1500?) to 1548B. Yes, I know where you got “14” (dst MAC, src MAC, type/len), but that’s not all the bits in a packet. I mean, … “kinda”. It’s all of L2 if you don’t count the FCS (CRC) or a 802.1q tag.
I got there by mistyping 1538 as 1548. I had two of my fingers smashed 3 days ago in a door (hinge side, hello Archimedes!) and typing has suffered as a result. Regarding 1514 MTU, I think that's a typical way to refer to a standard Ethernet (L2) MTU, but yes "MTU" by itself is always confusing given how varied OAM UIs used it. Thanks, Chris.
My typical understanding is 1500 byte MTU and a 1538B packet with all framing overhead, including SFD to IFG. For the life of me, I can’t get another 10. I could get 8 with 2 802.1q tags (Q-in-Q), but I can’t get 10. For Ethernet a 1500 byte frame turns into 1538 “on the wire” (1542 with 802.1q), but 12 bytes of that is “silence”. So perhaps you could educate? Jim (I’m also having trouble with GHz.. I mean, SFP+ direct attach is 10.3125 Gbit/s…. 10GBaseT is 4 lanes at 800 Msymbols/s, … )
signature.asc
Description: PGP signature
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#13166): https://lists.fd.io/g/vpp-dev/message/13166 Mute This Topic: https://lists.fd.io/mt/31821738/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
