Dmytro Shypovalov 🇺🇦 🇮🇪

TD release 1.7 is available! More options for path diversity, PCEP multipath, global adj SID support and more.

vegvisir.ie/2025/07/28/t...

Traffic Dictator v1.6 is available! A major PCEP rework (including dual-PCE redundancy), multi-domain policy improvements and some bug fixes. vegvisir.ie/2025/07/07/t...

Getting started with SR-TE never been easier: Traffic Dictator images are now available in a public Docker Hub repository: vegvisir.ie/2025/05/26/t...

TD version 1.5 is available! Controller redundancy and various quality of life improvements: vegvisir.ie/2025/05/08/t...

Generating an optimal segment list for SR-TE Segment Routing simplifies MPLS for the network operator – but not for the developer.

[Blog] How SR-TE implementations how generate optimal segment lists in any topology routingcraft.net/generating-a...

A guide showing how to deploy TD in a Docker container on IPInfusion OcNOS and establish BGP and PCEP sessions between TD and host system:

vegvisir.ie/2025/03/31/d...

TD release 1.4 is available! New feature - disjoint paths for traffic engineering and some bug fixes.

vegvisir.ie/2025/03/21/t...

TD release 1.3.1 is available, with some improvements and bug fixes.
vegvisir.ie/2025/02/21/t...

Loopback as a Service Methods of steering traffic into MPLS and Segment Routing LSP is one of the least standardized and most confusing parts of traffic engineering.

A new blog post about using BGP and multiple service loopbacks to steer traffic into MPLS-TE and SR-TE policies:

routingcraft.net/loopback-as-...

RFC8277 breaks backwards compatibility A quick public service announcement for anyone implementing BGP-LU or deploying it in a multivendor environment.

[Blog] RFC8277 breaks backwards compatibility routingcraft.net/rfc8277-brea...

Question to PCEP / Traffic engineering gurus - are there positive use cases of multivendor PCEP deployments? E.g. Cisco/Juniper/Nokia PCE with another vendor PCC.

RFCs are so vague like it's made on purpose for vendors to interpret everything in their own unique way.

I hate writing PCEP code so much. Just to generate one message, I need to check at least 5 different RFC. Even BGP-LS wasn't that bad.

Wonder why I keep saying PCEP is ridiculously overengineered?

A niche protocol doing one small thing has more IANA assigned parameters than BGP. Even the page is much bigger.

PCEP page references 45 RFC (vs 63 for BGP) but has more drafts so it's pretty close.

Simulated engine failure- glide approach on a short runway