4 Comments

Good posting as usual, but I do not agree with this comment: “Technological neutrality” advocates would have us ignore the the type of technology used and focus instead on the “maximum advertised speed”.

I think we can agree that advertised speeds can be an issue and for all technologies, especially fixed wireless but even including fiber. However, we can have tech neutral policy and focus on delivered or deliverable speeds, not just maximum advertised speeds. Thanks.

Expand full comment
author

Thanks Paul. Fair point. I agree that if we had "delivered or deliverable" throughputs that came from an established and reviewed process (it would have to be largely based on the technology, but also including other environmental factors) it would bring a lot of clarity to this issue.

Expand full comment

That's a hard thing to call. We should not focus exclusively on "maximum advertised speed" in part because it tends to be overstated, but more important because there are other metrics to consider, especially latency and reliability. The Broadband Forum has created a metric called a "Quality-of-Experience Index" that might be a good objective to focus on (I haven't gotten too deeply into it, so can't recommend it.)

All unlicensed FWA is not created alike. The latest technology, like the Tarana Wireless gear that NextLink is using for part of their RDOF build, is actually quite good, albeit expensive. So, I understand, is the unlicensed mmWave gear like Adtran/CCS and Terragraph. It shouldn't be lumped with WiMax or outdoor WiFi. I suspect that bad experience with WISPs using that technology and maybe lacking engineering expertise has poisoned a lot of people's opinions of unlicensed FWA.

Ditto DSL. Most of it is dismal. But there are pockets of VDSL that can easily clear the 25/3 bar, and even some that can do 100/10. There's not a lot of G.fast on this side of the Atlantic, but in parts of Europe, it is providing close to 1 Gbps downstream. Not to advise giving BEAD funding to Tier-3 telcos to upgrade their ADSL+ to bonded VDSL or G.fast. But it's hard to think of a principled techical reason to preclude it, either.

There's also the question of CBRS General Availability Licenses.

Wish I had time to get more deeply into this.

Expand full comment
author

A lot of great points here. I do a lot of work on working latency/responsiveness and agree that it is critical to the overall experience.

I've also heard good things about the Tarana gear. I wish between technology codes, propagation models, and some kind of agreed-upon translation to expected performance, we could produce a more accurate estimate of served/underserved/unserved.

On the Form 477 data there was a technology code that specified the type of DSL, and the Docsis version. I think that's the right direction.

Expand full comment