OK so I've been saying this for literally years and nobody listens but I am done being quiet about it. Token ring is dead.

Not "on its way out" dead. Not "declining market share" dead. Like actually dead and someone needs to tell the corpse. I have spent 20 years in computer engineering and I have watched more technology hype cycles than I can count and this one is special because it's the opposite of hype — it's a thing everyone knows is over and yet here we are still pretending.

The Technical Case

Ethernet won for boring, correct reasons. It's cheaper. The hardware is everywhere. The cabling is the same cabling everyone already has. You don't need a hub-ring topology and the failure modes it brings with it. You don't need to manage token rotation.

Switched Ethernet took the last real argument (collisions under load) and eliminated it. At that point Token Ring had nothing left but inertia and the sunk cost of existing IBM installations.

The Real Reason It Took This Long

The real reason Token Ring is still being argued about in 1999 — and I have had this argument in person, with real engineers, this year — is that the engineers who built their careers on it don't want to admit it's over. That's a human problem, not a technical one.

I don't say this to be mean. I say this because I have made the same mistake in other domains and the only cure is to say it out loud.

What You Should Actually Do

If you are still running Token Ring in production: you don't need to rip it out tomorrow. You need a migration plan. Start with the edge, move toward the core. Don't let anyone sell you on Token Ring upgrades as a bridge — there's no bridge, just more commitment to the wrong answer.

If you are evaluating a new installation and anyone brings up Token Ring: ask them why. The answer will tell you something important about how they reason about technology.

Ethernet. Get there.