iptrans

There are only about 100k ASNs. It would be fairly easy to solve the bootstrapping problem by just preloading all the keys.

Not really.

The parent to my original post identified a problem, but misdiagnosed the root cause and prescribed an incorrect solution.

As noted above a lot of poles are owned by power companies that do not offer broadband service. Furthermore exclusive franchises have not been permitted for a good while anymore.

Anybody with the correct permits can secure space on poles. This isn’t the issue. Part of the problems is how costs are assigned and how timely work progresses.

Who owns the pole isn’t the big issue.

> If only the first strand is expensive in a fiber deployment, why duplicate work by not installing residential fiber at the same time?

Because 5G is expected to be more profitable than fiber and deploying fiber would require more CAPEX.

There are a lot of (high level) resources online, but on a practical level you will have to deal with local permitting and right of way issues, so there really can’t be an ultimate mater plan.

Interestingly interconnection, or more precisely Internet capacity, isn’t really an issue. If you have enough subscribers, it is fairly straightforward to purchase wholesale bandwidth from one of the many commercial providers.

Nah. You can get a 4 port 10G switch for about $100 from a known vendor.

The US has a well developed and competitive national backbone market.

When building networks the geographical extremities of the nation do not matter, only the distance to the closest point of presence.

Europe and the US are not that dissimilar in geographical extent or population density, so neither of these can be used as convincing arguments for American exceptionalism.

The fact remains that the US is just that dysfunctional compared to Europe.

This appeal to American exceptionalism does not stand up to scrutiny.

A lot of European countries have far lower population densities that the US, but still better infrastructure and broadband.

The US just is that dysfunctional.

This isn’t about symmetric connections. It’s about Altice lowering the upstream from 35 Mbps to 5 Mbps.

Going from 35 Mbps to 5 Mbps is ridiculous. It effectively lowers the quality of the broadband by bottlenecking connection.

You need upstream bandwidth to use your downstream. Having only 5% upstream of your downstream means the connection is at risk of congestion merely from return packets when downloading.

Once you saturate your upstream your downstream will become laggy and slow down.

Even ignoring headline downstream speeds, 5 Mbps of upstream bandwidth is a piddling amount of bandwidth. It is trivial to exhaust that amount of bandwidth just by normal Internet usage and a video call.

The whole thing is just a money grab from Altice by forcing customers to upgrade to more expensive packages just to get working broadband equivalent to what they had before.

> It's cheaper per capita to roll out fibre to 50 million people than to 5 million;

This assumes an awful lot of things.

If the average number of linear feet constructed grows when going from 5 million to 50 million then the cost per capita isn’t going to decrease.

What going from 5 million to 50 million enables is averaging high cost builds over a larger subscriber base resulting in affordable prices for more people. Note that this will not however decrease costs overall.

The above is, in conciliatory terms, hyperbole.

In reductionist terms the real underlying reason for market failure is regulatory capture.

As to utility poles, a large number of poles are owned by power companies who by and large do not offer retail broadband services, so even here you are barking up the wrong tree.

That being said, there are issues with utility poles which need to be addressed. Efforts such as one touch make ready and the like are needed on a national level.

Very few fiber optic systems are run at capacity.

Upgrading bandwidth is therefore a matter of new optics and router cards, not new cables.

Furthermore new subsea cables bring down the unit cost of bandwidth.

Far more subsea cables have been decommissioned due to them not being cost effective anymore rather than not working.

Light travels more than a 100 miles per millisecond, so as long as the POP is within a few hundred miles you are good.

Tromboning local traffic isn’t much of an issue anyway since very little of your traffic is hyper local. Most of your traffic will be going to CDNs and other ISPs.

I assume the grandparents refers to the tromboning inherent in the NZ design.

This is a solved problem.

The municipality provides the fiber network. All ISPs rent dark fiber, local loops, wavelengths or bitstream access on equal open access terms.

All on equal footing and equal terms.

WiFi is really bad at anything but in home connectivity.

ISPs would more likely than not take exception to such scraping.

Even if they weren’t reporting accurate data, which they don’t always do.

WiFi is really poor as a last mile technology.

Muni wireless is generally a bad idea. It has been tried many times and failed.

What you want is a municipally owner fiber network.

What you really need to start a muni network is grassroots support. If the constituents want it it will be built.

You can start your own community ISP if you can sign up enough neighbors.

However, it’ll take years to build out and they payback will be far longer than two years.

There’s even funding to be had, but it’s no small undertaking.

Hit me up by email if you want more information. Contact in profile.

Alas, no such service exists.

Even the FCC can’t get proper broadband coverage data by the zip code, not to mention by address.

Starlink is only an option for a very, very small portion of the population.

The Starlink constellation has less total bandwidth than a single strand of fiber. It simply does not scale.

Starlink only has permits for about a million terminals or so and they must be very careful in where and how many customers they sign up so as to not overwhelm the system.

Elon himself has said that Starlink is only an option when you have no other options.

The same cable was used for cell service...

Less sexy headline: “Internet down in Canadian village due to poor planning and lack of proper redundancy”

Use the front page RSS feed.

Can somebody add [video] to the title?

Repeaters are designed to last for the lifetime of the cable plant. Design lifetimes are 25 years or so.

Repeater design is inherently very, very conservative because if the repeater fails, the cable fails. This results in an outage lasting days, if not weeks, as a cable ship is dispatched to the failure location.

The cable ship has to trawl for the cable and pull it up to the surface. Then the cable is cut and replaced with a new section that includes a new repeater to replace the failed one. Expensive.

Yes, not that there's a large difference. Best Buy has pretty large markups, especially on short CAT6 cables.

You can buy subsea cable for $10-$20 per meter.

EDIT: the cost depends on how many layers of armoring you require. Deep sea cable requires less, shallow sea cable more.

Fortunately you only need repeaters every 80 km or so, so you'd only need a bit over a hundred repeaters across the 9000 km span.

Repeaters aren't terrible expensive, so they only add a few million to the total cost.

10x is a fair estimate of cost vs regular armored fiber cable.

Source: I've laid subsea cable.

> how does the mesh connect back to the internet?

The same way any other network connects to the Internet, through an ISP.

Nominally you can be an ISP yourself, but you still need to buy capacity from other network providers.