San Jose's Google Fiber rollout is delayed to explores alternatives | Hacker News Reader

San Jose's Google Fiber rollout is delayed to explores alternatives

mercurynews.com

Verizon's costs to deploy FiOS, which is the same underlying tech as Google Fiber, between 2004 and 2010 in some MSA's approached a billion in capital alone. They ended up spending $23 billion in total capital, and only in places where they already had data centers every few miles, conduit, fiber, poles, field cabinets, franchise agreements, a local workforce, a large customer base, storefronts, etc. And that only includes new equipment and construction, not recurring operational costs. $23 billion is more than Google's entire capital budget for the past 2.5 years.

This is almost certainly the result of somebody at Google finally asking serious questions about the total cost, or at least being asked to cut a particularly large check for the first time.

I am also skeptical of Google Fiber's economics, but they have some business model hacks up their sleeve that FIOS doesn't: redlining and rallies. FIOS spent a lot of money to pass homes that will never subscribe and they had to do a lot of separate truck rolls to connect customers over time. Google Fiber is more like Kickstarter where they don't have to do any work unless the 'hood is guaranteed to be profitable, then they wire the whole place at once. Whether this business model is good for society is an open question.

Here in Nashville, the are rolling out access to only the richest communities first. But -- they are also about to wire some poor, subsidized Section 8 housing as well. It appears those of us in typical suburbia are likely last on the list, if we are on the list, for better or worse...

Hah. Well in Dallas, AT&T rolls out their fastest Gigabit speed (and most expensive offering) to the poor communities first. Because city council and politics. Makes sense.... the poor people, right after shopping with their food stamps and using their entire paycheck on rent can most definitely afford $130+/month internet...

FWIW, in Atlanta, signups just opened for residential service in some neighborhoods (I signed up today). So this is definitely not nationwide.

I just got the email. Too bad Brookhaven's availability is limited to apartments right now.

Looking forward to signing up but I am not sure how it works since I am off a street and not in a "neighborhood".

heck they tore up our yards in my sub, sent mail prior that we are getting fiber, only to now state there is no fiber in our area.

What a joke. So they're abandoning an already-finalized construction plan in favor of a return to the "service" that was https://en.wikipedia.org/wiki/Clearwire ?

Fixed wireless has been done poorly and it has been done well (Webpass). Fiber has also been done poorly (FIOS at >$2,000 per customer), so I don't know if historical anecdotes are useful here.

FIOS may have been expensive to roll out, but the quality of service was good.

> the quality of service was good.

But that doesn't make it a good business. And when it's not a good business, it gets scrapped.

Seems that Google was very premature about announcing these peninsula cities as rollout targets.

Wireless is better than nothing, but it will never compete with wired connection to the premises.

As an aside, I've always wondered if 5G or some other pending wireless tech could be used to one day replace Comcast...

5G is still a shared-medium carrier so it will get congested no matter how fast the total is - all it takes is a few bad or uneducated players spamming the wireless spectrum. Or just bad weather.

This is not wireless like 4g/5g or any cellular company you think of. This is point to point wireless they are talking about. Which can transmit incredibly high speeds and bandwidth.

Here's the idea, and is actually the current setup I have with a private company:

A p2p transmitter can broadcast to a lower tower that sits next to multiple neighborhoods. Then from that tower, lines can be run to individual houses and/or other small receivers like on people's rooftops.

With this system, they can literally cover vast areas, with just as fast speeds, in a fraction of the costs and time.

I have this setup right now, pointing my p2p device right at the top of a building downtown (not in San Jose.). It is crazy fast and even during heavy storms, there is no disruption.

San Jose, I wouldn't complain just yet, you might actually get your 1 Gbps + speeds faster than you thought despite this "construction delay".

How does the latency compare to a wired connection? Is it viable for gaming and the like?

Often very low. Urban 24Ghz/60Ghz point to point connections are usually 2ms or less per hop.

My house (rural small town) is 4x5Ghz & 1x11Ghz hops away from my datacenter and I get ~15-20ms ping times to Google, League of Legends, etc). At various times I've had Frontier VDSL2 and Charter DOCSIS and both had higher latency.

My connection point is about 1.5 miles away. I get about 4-6ms latency on my end device. That's going through my router, wireless access point, to my device.

and it's also very consistent. It's not dealing with the million redirects and other traffic that going through the phone line or cable line will have.

In very heavy rain and hind wind, my direct TV dish was out at the start of the storm, my internet p2p connection never lost signal, didn't even slow down.

Oh wow, this is extremely interesting. Hopefully Google rolls their connections out soon.

Yeah, wireless is super fast, satellite is the one with latency (and usually then due to dial-up uplink and just the distance of the sats). Most p2p links are under 50 miles, and in this case probably under 5 miles. The latency from the wireless link will be a negligible (for those not in HFT) overhead on the order of a few ms.

SJ, sure -- though for other peninsula cities (Mountain View, Palo Alto) I'm curious as to how well this type of solution would fare considering the lack of high rooftops and the lack of space/pervasive NIMBY attitude that happens when someone wants to build something like a tower for p2p transmission.

I was really stoked to see Fiber coming to my house in the near future, but this doesn't look good for places like MV/PA.

I'd love to hear more about how this is done. Care to share links, websites, etc? Or to write me, dwanjiru at Google's email service about it?

WiLine (http://www.wiline.com/) in San Francisco is one such high-speed, point-to-point wireless provider. Up to 10 Gbps symmetrical bandwidth and sub-10ms latency.

Google Fiber just acquired WebPass, which is an ISP in San Francisco and Oakland that already uses point-to-point wireless in a bunch of high-rises there.

Having fiber optic to the building (ala AT&T Gigapower) isn't all that useful if the entire neighborhood is oversubscribed 50:1 -- during PM rush, whether you use FTTH or point-to-point wireless, you're going to see speeds top out at ~20-200 Mbps on a "gigabit" link no matter what.

As a Webpass customer it definitely competes. I frequently get 500Mbs during prime time hours. Webpass is microwave wireless to the building, then the "last yards" to the unit is gigabit ethernet.

Google actually got that far and then backed out? That's incompetent management.

Palo Alto has mostly underground utilities. No poles. Did someone at Google not know this?

I'm curious how webpass is achieving this type of coverage. I live in a pretty small town and our one Wisp has saturated nearly every channel of 900mhz, 2.4ghz and 5ghz. I work with the local radio station frequently that leases tower space to them. I know for a fact they have no spectrum left for unlicensed backhauls in those frequencies. It seems like the problem would be greatly amplified in urban areas unless they used licensed backhauls.

I put some Google Fiber build, sales and financial data into a basic model to create forecast estimates Gfiber.us

Why is Silicon Valley of all places struggling to get reasonably fast Internet?

Most of these cities already have gigabit fiber through AT&T.

Title's misleading -- it's only delayed in one area (Silicon Valley).

Thanks. We updated the title, which originally omitted “San Jose”.

They've also put Palo Alto (which has already had the fiber backbone in place for decades) and Mountain View on hold.

The Merc changed their headline after publishing - see the URL, and how it was originally syndicated: http://www.topix.com/com/google/2016/08/google-fibers-silico...

Also, in the second paragraph, it specifically calls out to other SV cities where delays are confirmed - Palo Alto and Mountain View.

Also delayed in Portland[1]

[1] http://www.oregonlive.com/silicon-forest/index.ssf/2016/07/g...

Reply on news.ycombinator.com