ISIS is normalised and preferred. Segment routing would be great for proper traffic engineering and routing symmetry in large scale ISPs. Even without MPLS/VXLAN.
I hope MikroTik supports both soon.
Probably because the new routing engine in RouterOS v7 is technically more advanced than FRR.
It’s possible. I’ve not performed a full 1:1 feature parity analysis between them. That being said, I didn’t make the assertion that one is superior to the other. I was making the point that MikroTik could have had an opportunity to start with an existing routing stack that’s proven in production, and all the features thereof. They could also have moved the state of FRR forward with their own contributions back to the community; but they didn’t, and that’s a decision I can respect.
I am not particularly biased either way, as it’s more of a “use the tool that fits the job” question for me. RouterOS is good enough for most purposes; when I need a feature RouterOS doesn’t have, usually I can find it in FRR (segment routing just happens to be the topic of this thread, so I’ll use that as an example; other (current) examples include BGP ORR, BGP LS, BMP (partially supported by RouterOS, IIRC?), BGP Flowspec, etc).
We have ISIS as of RouterOS 7.13, so that’s cool. I’m sure MikroTik will consider some of the more complex use-cases enabled by the various extensions to BGP that have come out of the IETF eventually. So, to me it’s only a matter of time.
Anyway, all the best to you and yours, mate.
IS-IS is here, +1 for SR-MPLS, that would be great to have.
+1 upvote here for SR-MPLS and SRv6!!!
Why use something open source when the manufacturer has their engineers working on its implementations.
Anyone in MikroTik able to confirm or deny if Segment Routing is actually being considered (seriously, not just on the drawing board for the past 10 years) or worked on?
It’s the single most important feature for my use case, and should be one of the top priorities for any ISP that starts scaling out to more than a few dozen sites with redundant paths etc
Having the ability to direct traffic through your entire topology, and not just next-hop is vital to effectively utilizing backup/redundant paths and scaling up. It just can’t be done with OSPF or IS-IS on its own