Yeah, but you had some rift hacked I remember. In case it’s mooted, we’ll drop you off this list
Juniper Business Use Only
From: wet sh1t <>
Date: Friday, 29 September 2023 at 21:26
To: Antoni Przygienda <>, Jeffrey (Zhaohui) Zhang <>, Jeff Tantsura <>, <>, <>, <>,
Mohácsi János <>, Farkas István <>, Csendes József <>
Cc: Jordan Head <>, <>, Bruno Rijsman <>, Zoltan Salamon (zosalamo) <>, Alexander Gall <>, Simon Leinen <>,
Visky Balázs <>, Csendes József <>, netadmin <>, Közbeszerzés (KIFU) <>, Dr. Mate Ferencne <>, <>
Subject: Re: RIFT interop in IETF118 hackathon?
[External Email. Be cautious of content]
hihi bruno, hi zte guys, nice 2 meet you, again..... :)
(evening tony @ junos, zosa @ cisco, alex & simon @ zurich, *kifu*... :)
freertr.org is definitely a different codebases, its whitebox router written for 10+ years now
and it recently got/interopped/demoted a p4 programmable whitebox dataplane (read, switch) based
around the the 12tbps intel barefoot forwarder, intel dpdk.org (100gbps smartnics included like cisco
fd.io, the asr1k-x-esp200 onwards, the _real_ vector processor base boxes, nowadays the cat8k boxes to
name the loved-hated-beasts...)
but, for just this n-th routing protocol in freerouter, even the legacy openvswitch
(openflow) exporter can do the tricks to be "in-hardware" on the 3rd party "firewall"
devices like the corsa programmables (10y+ story) dunnowhat i tested in the geant.org
testbeds in an earlier stages of the gn5-1 project...
this year we get verified that the cisco silicon one is _publicly_/onlist- verified that p4
programmable as rumors said earlier;
https://urldefense.com/v3/__https://twitter.com/CiscoIOSXR/status/1466006474698731529__;!!NEt6yMaO-gk!ELI09iNGpEDocSkiePhVxp54CQWXphacIKg-Y4JYeumZLvrlVdNVZFX2_iP1qQduDtAoLPmyQg$
https://urldefense.com/v3/__https://www.cisco.com/c/en/us/solutions/collateral/silicon-one/datasheet-c78-744312.html__;!!NEt6yMaO-gk!ELI09iNGpEDocSkiePhVxp54CQWXphacIKg-Y4JYeumZLvrlVdNVZFX2_iP1qQduDtANsatpfw$
it was on my personal bucket list for years and thanks to the cern/switch contacts it fulfilled...
moreover there are youtube links flying around that then cisco can install m%-azure-sonic layer2 to these... :)
next years target of ours is the broadcom nplang.org and the surroundings like the ncs5k5
if a whitebox vendor jumps to the fun-wagons...
once you bumped up anything and there is a _significant_ change in the tlv structures that makes
freerouter<--->python implementation incompatible, just let us know plus the suggested diffs and
moreover the reasonings, this stuff performed interop on a virtual ietf already so we also would
be interested how why etc this gets broken as it happened earlier to the ongoing draft bgp-ct, or
the bit indexed explicit replication..... below you can find some proof of the rift ipv4 ipv6
in production configurable, downloading the codes from github then ./tw.sh rout-rift01 capture eth1
can produce you a self-test with a pcap output in a docker/vm/baremetal/whatever...
ps: b4 hitting reply-all please remove kozbesz* && *nop, as its a lawyers/accountants lists, so not technicians....
happy weekends,
cs
mchome#
mchome#
mchome#show clock
machine: 2023-09-29 21:17:40
email: Fri, 29 Sep 2023 21:17:40 +0200
zone: CET diff: 00:00:00.000
mchome#
mchome#
mchome#attach vdc wifi
escape character is ascii0x03.
wifi#
wifi#
wifi#
wifi#configure
wifi(cfg)#
wifi(cfg)#
wifi(cfg)#router ?
aggreg4 - auto aggregate creator to configure
aggreg6 - auto aggregate creator to configure
babel4 - babel routing protocol to configure
babel6 - babel routing protocol to configure
bgp4 - border gateway protocol to configure
bgp6 - border gateway protocol to configure
blackhole4 - blackhole collector to configure
blackhole6 - blackhole collector to configure
deaggr4 - deaggregate creator to configure
deaggr6 - deaggregate creator to configure
download4 - route download to configure
download6 - route download to configure
eigrp4 - enhanced interior gateway routing protocol to configure
eigrp6 - enhanced interior gateway routing protocol to configure
flowspec4 - flowspec to flowspec rewriter to configure
flowspec6 - flowspec to flowspec rewriter to configure
ghosthunt4 - ghost/zombie route hunter to configure
ghosthunt6 - ghost/zombie route hunter to configure
isis4 - intermediate system intermediate system to configure
isis6 - intermediate system intermediate system to configure
logger4 - route logger to configure
logger6 - route logger to configure
lsrp4 - link state routing protocol to configure
lsrp6 - link state routing protocol to configure
mobile4 - mobile route creator to configure
mobile6 - mobile route creator to configure
msdp4 - multicast source discovery protocol to configure
msdp6 - multicast source discovery protocol to configure
olsr4 - optimized link state routing protocol to configure
olsr6 - optimized link state routing protocol to configure
ospf4 - open shortest path first to configure
ospf6 - open shortest path first to configure
pvrp4 - path vector routing protocol to configure
pvrp6 - path vector routing protocol to configure
rift4 - routing in fat trees to configure
rift6 - routing in fat trees to configure
rip4 - routing information protocol to configure
rip6 - routing information protocol to configure
uni2flow4 - unicast to flowspec converter to configure
uni2flow6 - unicast to flowspec converter to configure
uni2multi4 - unicast to multicast converter to configure
uni2multi6 - unicast to multicast converter to configure
wifi(cfg)#router
On 9/29/23 20:12, Antoni Przygienda wrote:
> So (especially at Sandy) the requirements would be to have suppor for the .yaml file format and run either the odd or even nodes. This can be easily tested against Bruno ;s
> implementation/framework that can be easily extended to test that.
>
> Once we have that in place we can test any matrix of 2 implementations.
>
> I ll update Bruno s open source to newest spec/schema so we re all in line.
>
> Goals AFAIS
>
> 1. Bring up adjacency v4 (would be good to do v6 as well) with fixed level config
> 2. Bring up adjacencies with ZTP, in one case ToF node, other case without ToF flag
> 3. Exchange TIEs until databases coherent between the nodes
> 4. Verify routing tables (if implementation allows)
> 5. Test leaf flag if implementation allows
>
> Once that is all in place we can go and start breaking links and test disaggregation (Bruno s framework has that included and implementation needs to pass that first against his
> testframework before other combinations I would say)
>
> Folks can chime in with what they want to try, what they have to test
>
> * Tony
>
> *
> *
>
> *
>
> Juniper Business Use Only
>
> From: *Jeffrey (Zhaohui) Zhang <>
> *Date: *Friday, 29 September 2023 at 17:48
> *To: *Jeff Tantsura <>
> *Cc: *Antoni Przygienda <>, Jordan Head <>, <>
> *Subject: *RE: RIFT interop in IETF118 hackathon?
>
> You re right.
>
> Juniper Business Use Only
>
> *From:*Jeff Tantsura <>
> *Sent:* Friday, September 29, 2023 11:40 AM
> *To:* Jeffrey (Zhaohui) Zhang <>
> *Cc:* Antoni Przygienda <>; Jordan Head <>; <>
> *Subject:* Re: RIFT interop in IETF118 hackathon?
>
> *[External Email. Be cautious of content]*
>
> No, FreeRTR is a different code base, sorry, I didn t realize ZTE is in.
>
> Cheers,
>
> Jeff
>
>
>
>
> On Sep 29, 2023, at 08:33, Jeffrey (Zhaohui) Zhang < <>> wrote:
>
>
>
> FreeRTR is Bruno?
>
> ZTE is all in.
>
> Jeffrey
>
> Juniper Business Use Only
>
> *From:*Jeff Tantsura < <>>
> *Sent:* Friday, September 29, 2023 11:32 AM
> *To:* Antoni Przygienda < <>>
> *Cc:* Jeffrey (Zhaohui) Zhang < <>>; Jordan Head < <>>;
> <>< <>>
> *Subject:* Re: RIFT interop in IETF118 hackathon?
>
> *[External Email. Be cautious of content]*
>
> So, practically would be Junos + Bruno + FreeRTR?
>
> Cheers,
>
> Jeff
>
>
>
>
>
> On Sep 29, 2023, at 00:54, Antoni Przygienda < <>> wrote:
>
>
>
> And I would tell them to plug into Bruno s framework where e thing gets tested against Bruno framework. I.e. an implementation has to run bunch nodes and connect based on
> yaml on either red or black nodes in the yaml
>
> This will allow to run any 2 implementaton combination against each other
>
> Juniper Business Use Only
>
> *From: *Jeffrey (Zhaohui) Zhang < <>>
> *Date: *Friday, 29 September 2023 at 05:19
> *To: *'Jeff Tantsura' < <>>, Antoni Przygienda < <>>, Jordan Head
> < <>>, <>< <>>
> *Subject: *RIFT interop in IETF118 hackathon?
>
> Hi,
>
> Did we talk about that? We may have but I forgot about that entirely, though we did give Juniper's RIFT image to ZTE for interop testing.
>
> Besides Juniper, ZTE and FreeRTR, who else has the implementation?
>
> Jeffrey
>
> Juniper Business Use Only
>
|