Re: How does FreeBSD expect to compete in a DPDK/VPP world?

I think there is a mix of things here.

For the VXLAN or MPLS this depends on the role of BSD.

For DC with leaf and spine deployment, nowadays VXLAN is good enough, as 
we can terminate the VXLAN on FreeBSD + Bhyve.

Now if you are in a transport network or need to terminate services 
(L2VPN, L3VPN, VPLS ,EVPNoMPLS) then MPLS and SR-MPLS are required for 
the data-plane and RSVP, BGP-LU and ISIS/OSPF with SR extension for the 
control plane. Some networks are still running LDP but nobody will 
deploy a greenfield on or a modern network on it.

On the DKDP and VPP, the intention is to bypass the kernel, so i don't 
see why FreeBSD itself should come with a solution for that apart from 
making sure DPDK and VPP work on FreeBSD. The last time i spoke with 
somebody from Intel, the response was that DPDK was is good shape (For 
FreeBSD), but also I remember some email on this list saying that some 
things are missing. Not sure what the current status for VPP.

Regards.

Santi


On 7/21/22 19:38, sthaug@nethelp.no wrote:
>>> Not that I work at an ISP or tech company in a networking role, I don't.
>>> Heck, adding even MPLS support has been on my bucket list for a while, but
>>> am too lazy to get started. I do want to move to a networking-based role at
>>> $DAYJOB, but we'll see about that.
>> MPLS is outdated (performance-wise) and (functionally) replaced by vxLan, or
>> do I miss something?
> No idea what you mean by outdated (performance-wise). Yes, I know the
> initial claims for MPLS touted faster packet switching times - but
> that hasn't been relevant for many years.
>
> MPLS is about functionality. Yes, there is some functionality overlap
> with VXLAN - but the overlap is certainly not complete. Personally,
> I'd love to see an MPLS implementation in FreeBSD.
>
> Steinar Haug, Nethelp consulting, sthaug@nethelp.no
>