Hacker News new | past | comments | ask | show | jobs | submit login
Google Fiber agrees to acquire Webpass (webpass.net)
304 points by dweekly on June 22, 2016 | hide | past | web | favorite | 128 comments

Huge fan of Webpass. 1 Gbps of IPv4/6 with great customer service for $55 a month. It can't be beat, if it's available to you.

I hope this is a win for the Webpass employees, in the short and long term. Thanks for all the bits!

We once spent a year in a building that had Webpass. Ever since then we've moved several times, and every time we ask if they have Webpass in the building. Fantastic service.

$60/month now but I agree with all your points.

My only complaint is that for home service there is no way to get a public IP address, so you've got to get an external server if you want to access your home computers from elsewhere on the Internet.

Yeah they ran out of v4 addresses a little while back, but you can always get an IPv6 address from them. To avoid having most people get confused by IPv6 though they seem to have some form of layered IPv4 NAT (one IP per building or something, and then your "own" IP), which in a different building I discovered and found more confusing than "oh look, just straightforward v6 only".

> "oh look, just straightforward v6 only"

An ISP that doesn't provide at least outbound IPv4 connectivity would rapidly lose all of its customers.

IPv6-only with NAT64 is doable when you control the set of attached devices; see T-Mobile US, with phones that implement 464XLAT.

Earlier in this thread, it was mentioned they are offering IPv6. I'd expect those addresses to be public at least?

I wasn't aware of that. I'll have to look into it.

Internal IPv6 addresses are publicly addressable on Webpass but you may have to enable it on your router.

If you have terminal access to your router and it supports iptables, you can manually enable forwarding for a specific port like this:

ip6tables -I INPUT -p tcp --dport 5005 -j ACCEPT ip6tables -I INPUT -p tcp --dport 5005 -j ACCEPT ip6tables -I FORWARD -p tcp --dport 5005 -j ACCEPT ip6tables -I FORWARD -p tcp --dport 5005 -j ACCEPT ip6tables -I FORWARD -m state --state ESTABLISHED,RELATED -j ACCEPT

Note that if you're using this for Api callbacks like AWS Lamda, they do not support IPv6 and then your best bet is a VPN tunneled to your local device.

Good luck!

What's the use case for using port forwarding with IPv6 rather than just allowing the traffic to the machine in question?

I believe the above ip6tables rules already do what you describe; the only target is "ACCEPT", with no packet mangling.

Do you only need access for yourself? If so, you should really check out ZeroTier. I use it on my Android phone and Linux boxes and have been really happy with it.

This; it's awesome!

That still seems pretty expensive. In Romania you get the same service (1 Gbps with IPv4/6) for less than $12 (45 RON).

That's pretty cheap but you have to account for purchase price parity. Local goods are more expensive in America than Romania. By about 2X. That includes cost of local labor.

For example, the average Romanian wage is like 8,000 dollars a year. But that 8,000 dollars goes a lot further.

Internet is absurdly expensive in the US for what you're getting. For instance I pay $15/month for 2mbps.

What does the price of something in Romania have to do with the price of something in the US?

Is that a good deal in the US ?

Wow, in France for $40 we got:

- 12Gbps (100G Fiber at no additional cost in supported area)

- Free landline call to a 100 countries and free cell calls to France


- and a media center with recording capabilities, a blue ray player, public FTP and network harddrive

- for 3 more $, you get a cell phone line with 2 hours of communication and unlimited text, or for $20 unlimited calls and 4G.

And people are still complaining...

I think you are confusing Megabits and Gigabits.

100G fiber? I had no idea France was running 4x EDR Infiniband to their citizens' homes.

I demand only the finest quality 16QAM low FEC coherent single wavelength 100 Gbps interface for my condo, doesn't everyone have an ASR9010 with $50,000 100GbE linecard in their closet???

Do you have some proof of that? I know you can commonly get multi-gig connections in some countries, but France was next on that list that I know of.

Today I have 11Mbs internet in England. Before today it was the best I could get but today I was made up that my exchange just went live with 76Mbs fibre.

You guys kinda rained on my parade :(

I get 50Mpbs here with Virgin Media - pretty happy about it :D

12Gbps? Seriously?

Has to be wrong, no end user in France is on 100Gbs without ridiculously prohibitive costs

Dear Google: If you fuck up Webpass, you're dead to me.

p.s. I'm watching you: http://i.imgur.com/rS9LGZ6.png

What tool are you using there?

Just a little cron job shell script that dumps the results from a command-line speedtest.net client into Librato Metrics.

Wow, unexpected. I used Webpass when I was in SF and was very happy with the quality of their service, so them getting more capital to expand is a great development.

Yep, WebPass was by far and away the best ISP I've ever had. They actually emailed me in advance about planned service interruptions! Wow!!!

Interesting.. I've been very happy with Webpass for the past 8 or 9 years (though my building's crappy wiring can "only" support 100Mbit).

A Webpass engineer I spoke to last week mentioned they're rolling out fiber to the building to provide redundancy for the wireless mesh. Surprised me as I could count on my fingers the number of outages webpass has had since I started using them. Perhaps it'll replace the wireless portion in the longer term.

It might look like a mesh on a topology diagram but believe me, it has nothing in common with wireless "mesh" networking... I don't have insight into the exact configuration of webpass' routers but they're undoubtedly using fairly ordinary BGP and OSPF between nodes. The wireless links are exclusively high capacity point-to-point trunks.

It is not "mesh wireless" gear, but it is a mesh at the routing layer (ECMP). My old building had PtP wireless to multiple other buildings.

Edit: I always speculated the difference between 100Mbps and 200Mbps on-net buildings was if that particular location was multihomed.

Former Webpass employee here. Was one of the first software eng. hires. This is a great example of a bootstrapped business focused on customer experience that built up with no outside funding (that I know of). Congrats Charles & team.

Great work. I was a customer at 673 W Brannan in SF back in '07. Best residential internet ever.

I'm at 673 for the last few years, it's still great!

The idea of starting an ISP without outside funding (unless you're a millionaire already) is pretty astonishing. Been researching the possibility of this myself recently and costs of infrastructure are simple astronomical.

Going wireless is a lot cheaper. But you are right. They had to have had either founders capital or they took bank loans after they signed contracts.

I use Webpass. Get 600mbps on my wifi AC router. It's amazing and I'm super happy with it. Hoping with google fiber speed gets even faster.

How much are you paying ?

I paid $550 for a year up front. It's ~$55 per month on a monthly plan.

Damn. I pay windstream about $65/month for 10mbit down/768kbps up. (They're the only wired provider where I live, and they know it.)

Cool! What router/access point is that?

While I worry about large carriers absorbing small ones, this could be an excellent big step in real competition for bay area internet access.

If you listen, you can hear the sound San Francisco rental managers rejoicing they can advertise they have Google Fiber.

I have webpass with 1gb for 550 a year. I've had it for 7 years or so.

The bummer is- it's actually cheaper than Google charges in other cities- hopefully the price does not go up.

Webpass does not serve houses or smaller and/or older apartment buildings. This substantially lowers their install costs per customer relative to business models, like Fiber, that serve almost everyone n an area.

It's a brilliant business model: expand slowly based on demand only to dwelling units dense enough to be economically viable.

It definitely is and a bunch of small companies are doing an amazing job of it (Seattle has three such companies, for example).

It is also one big reason why I really want municipal fiber. There are a lot more "houses or smaller and/or older apartment buildings" than there are new condo or high-end apartment buildings. It would be nice to have some competition in those dwellings alongside the cable company, usually with caps, and (maybe) the telephone company, usually DSL with very slow speeds.

Municipal fiber wouldn't be "competition" in any meaningful sense of the term. The Chatanooga rollout cost about $5,000 per subscriber and was heavily supported by cross-subsidies from electricity ratepayers. I'd support municipal fiber as a complete government monopoly in places that have shown themselves capable of operating public infrastructure (e.g. NYC). But billing it as "competition" is disingenuous. You can't meaningfully "compete" with an entity that gets non-market revenue.

I have Verizon FiOS. Installation of it was subsidized by targeted tax breaks, use of government owned right of ways, and laws which substantially limited their liability for property damaged done during installation. I'm also pretty sure Verizon used funds gained from their other services to pay for the installation.

In most places FiOS deployment involves huge taxes, such as franchise fees and forced build-out into neighborhoods that will be unprofitable. (Tax breaks, to the extent they exist, are often carrots for the latter). Also, in many (most?) places Verizon pays the utility pole or conduit owner (usually the power company) for access to those rights of way. In Virginia, for example, you'll get a line-item on your bill for the right-of-way fee.

And in many places there is no forced build-out. The county I live in is only partially served by either FIOS or the cable company. The more rural part of the county is dependent on expensive WISP service because neither has been forced to build out into unprofitable neighborhoods.

The right-of-way fee covers costs for use of poles and conduits but also municipal right of way usage fees. The municipal right of way fees are much less than it would cost Verizon to otherwise negotiate and purchase right of ways from private property owners.

In most places telephone deployment involved huge taxes and forced build-out into counties that would be unprofitable. Also, the telephone company paid non-discriminatory fees to the pole/conduit owner.

This was often paid for / subsidized by the Universal Service Fund, because telephone service was seen as something absolutely necessary for modern life. Back in 1934.

ISPs, unfortunately, are not able to be subsidized by the Universal Service Fund. AFAIK.

Note that the Universal Service Fund is funded by a tax on telecommunications service. So a big company like Bell Atlantic (Verizon's predecessor) never received a net subsidy--the government just forced it to take money from some of its subscribers to offer below-cost service to other subscribers.

That said, the USF is a big step up from the "forced build-out" into unprofitable areas, which is what used to happen with telephone, and still happens with broadband. It highly distorts the market. E.g. if build-out requirements were imposed on Webpass, it'd basically ban their business model.

Fellow Seattleite - who is doing this here? I have a personal interest... :-)

This is exciting news! In Chicago, I'm currently living in a high rise where the fastest plan I can get is 24 Mbps for around $80/mo. My high rise has an exclusivity contract with AT&T U-Verse. I've spoken with AT&T reps and they can't offer any higher speeds. I also talked to property management and they said there's nothing they can do for me. They are locked into an exclusivity contract with AT&T for the wires in the building.

Does anyone have experience dealing with properties who claim to have exclusivity contracts? I talked to people at Webpass, and they've stated it is available in my area. They'd come in and set everything up free of charge. I don't see the downside for my building to allow Webpass to come in. I do know the FCC has regulations about exclusivity contracts with video providers, but I couldn't find any documents on things like internet. It seems like my only option is to find a place that does not have exclusivity contracts with providers.

> I can get is 24 Mbps for around $80/mo

Bucharest internet pricing is 10$ / 1gbps, so that would come at about 320x price ratio, and still 5x cheaper than Google Fiber.

Funny thing, the ISP I am getting this from set the price for 100mbps and 300mbps the same, 7$/month. As if a 200mbps difference is trivial.

Your property management company is likely receiving a significant sum of money from ATT to keep it this way. Unless you have access to their contract with ATT, you're SoL.

Hopefully this will speed up Fiber's rollout in San Francisco. I remember Google's announcement regarding Fiber in SF a few months back. This acquisition is a no-brainer for both companies!

God, I hope the support doesn't end up like most other Google product's support. I have webpass and my heart dropped. My project fi experience has been bad enough already.

I work on Google Fiber. We have pretty good customer support; call us 24/7, someone in the US will help you with pretty much anything you want that's Fiber-related (WiFi too slow? TV remote not working? That sort of thing.)

I just hope the commitment to good service that Google Fiber has started with isn't trashed by foolish executives.

The state of customer dis-service in the USA is so strong that it seems pretty likely any company that has good service for awhile will shift resources into bigger paychecks for executives and fewer resources for customer support. I hope Google Fiber avoids this situation. Some companies do but not very many: Trader Joe's, Costco...

I think good customer service is pretty much essential to running a service like Fiber, because there is no code that I can write that's going to smooth over the fact that we killed your lawn while running the fiber through your yard. It happens :/

Yeah we have you in Provo. Awesome service!

But can you tell me when the Stone Oak area of San Antonio will get fiber? Sorry, but living with TWC is a bear, and am hoping an actual date that I can add to the calendar of when to cancel would make it more bearable.

While the support will almost certainly get worse, the good news is that Google some of the most rigorous SRE practices in the world.

What issues have you had with project fi? I switched to it about 6 months ago and I haven't had any problems at all.

The issues are rather small. The support is what's bad. They expect you to contact them endlessly, through multiple mediums, to resolve the most mundane of issues.

I was a happy Webpass customer for about 3 years before moving to an area not served by them. I would be curious to see if they are folded into Fiber or if they are kept as a separate brand. As it currently stands Webpass has much stricter limits on who they are willing to serve. Last time I checked they required that all customers lived in multi unit buildings that were built within the last 20 years. I imagine this was meant to keep their installation costs low while hopefully benefiting from economies of scale. That makes sense for a new niche ISP, but Google has the resources and motivation to abandon those restrictions.

Has Google Fiber sold fixed wireless under the Fiber brand before? I think they would have to keep a separate brand at least until they can get all of the many Webpass fixed wireless subscribers switched over to fiber.

I hope this reduces the cost of their business offerings. I love the service, and the home offering is the deal of the century, but $500/month for 50Mb/sec symmetric or $250/month for 20Mb/sec symmetric is hard to swallow for our business (or any other).

They offer faster service (gigabit and probably more) but what startup can pay 10k/month for their business internet?

What's worse is monkeybrains and webpass should seemingly compete on price for business offerings but it seems like they actually collude. Their pricing for businesses is exactly the same.

With plenty of venture capital, $10,000 per month is not the greatest expense by far, in San Francisco. That’s 800 employees on Slack Plus, or enough space for about 20 employees with an open floor office and no meeting rooms, or less than 1/10 of the salary of an average software engineer.

Certainly, you should try to eliminate unnecessary expenses. That’s one of the appeals of cloud services: You no longer need to buy your own servers and get professional Internet service to reach the whole world. Nor do you need to spend the time running all that stuff, which can be significant.

If you do need fast Internet for your startup, have you considered Sonic.net? They have gigabit fiber only in the Sunset district of San Francisco and a few other neighborhoods in the Bay Area, but their price is only $90 for a small business, or $40 per employee when you have 5 or more “work stations.” That’s 250 work stations before you reach $10k/month, and by then probably you have enough money and need for room to pay for an actual SoMa office.

In some buildings, our (MonkeyBrains') pricing is less. Higher density = lower rates. The base rate covers hardware, wages, 100% employee health care, other infrastructure, and support. I imagine the costs between the two companies are similar.

I'm curious, why is it that business plans are so much more expensive than residential ones?

I think they expect businesses to use more traffic, and to require more expense to support.

Practically all Internet service is oversubscribed to some extent. The uplink is maybe 1 gigabit, but the downlink that has been sold is maybe 30 gigabits to 200 gigabits, when you add up all the customers. It is actually a feature of the Internet, that you don’t use any bandwidth when you don’t send anything, and you have the full bandwidth available to you when nobody else is sending anything. You can’t afford high-speed Internet if you had to pay for all the bandwidth yourself.

The problem comes if you use the bandwidth a lot. For most of the day, home service is idle, with a peak in the evening when everybody is watching YouTube and Netflix at the same time. And that traffic is mostly downstream, so a big company like Comcast can receive it for free, or even charge the source for “transit.” Businesses peak at business hours for their time zone, and a lot of businesses have a significant amount of upload traffic. If your company successfully expands overseas, then it will be uploading all the time. In theory, business plans cannot be oversubscribed as much.

If your web site doesn’t use HTTPS, then the big ISPs also tend to do deep packet inspection, and some of them inject their own ads. Fun times.

Of course, there are also captive audiences and rent-seeking.

Same reason a 64GB iPhone is so much more than a 16GB one.

Price discrimination. Businesses tend to be more able to pay than home users.

if by discrimination you mean guarantees versus best-efforts. business class services, speaking generally, are less or not at all over subscribed while providing contractual guarantees to service performance and problem resolution.

residential is best effort. they'll do their best to keep it up and give you the speeds you pay for, but if it doesn't work out that way sometimes, well, they're sorry about that.

Price discrimination is an economic concept - charging different prices for very similar products or services to get the value that each customer is willing to pay. That is often difficult and what normally happens is each customer is charged the same price.

Actually those are the rates for the best effort class of service. The business class service where you have guarantees is even more expensive.

That plus one or more of: SLAs, guaranteed minimum bandwidth, much higher upstream, better tech support, etc.

And static IPs, and a lack of language in the contract stating that servers cannot be hosted on it, and so on and so forth.

I'm pumped. I live in a Webpass market so hopefully Comcast starts doing their damn job now that it's a Google market.

So will Webpass be faster now?

My webpass is already 1 Gbps symmetric. If yours is less, it's a per-building upgrade (so maybe with additional capital they'd do that in more places) and may not be feasible.

Disclosure: I work at Google (but not on Fiber).

It's more the interconnection with the rest of the Internet than the building uplink... :)

My building is on 500mbit Webpass and things get pretty dire at peak times.

me too: Argenta.

Coming from australia and not familiar at all with Webpass... is the latency low? ie. could people use this for gaming internet connections?

As someone who used to live in Sydney and now in SF with Webpass... I can tell you that Webpass would be what a proper NBN rollout would look like. It's kind of amazing. It only costs $60/month and you just get an ethernet drop.

No login, no metering, no shaping.

Example latency:

  admin@RT-AC5300-E6A0:/tmp/home/root# ping -c 5
  PING ( 56 data bytes
  64 bytes from seq=0 ttl=55 time=3.526 ms
  64 bytes from seq=1 ttl=55 time=3.559 ms
  64 bytes from seq=2 ttl=55 time=3.547 ms
  64 bytes from seq=3 ttl=55 time=3.518 ms
  64 bytes from seq=4 ttl=55 time=3.584 ms

  --- ping statistics ---
  5 packets transmitted, 5 packets received, 0% packet loss
  round-trip min/avg/max = 3.518/3.546/3.584 ms

And no public IP, which was really disappointing when I switched from AT&T Uverse.

I'm always intensely curious what residential (and these days, even business) customers want with a dedicated, publicly routable IPv4 address.

why were you disappointed?

I wouldn't use it for anything too resource-intensive or important. It's just nice to have my home development environment available over SSH for the rare coding emergency. I shouldn't exaggerate how disappointed I am. It's obviously not very crucial to me, since it would be easy enough to set up a cheap VPN for tunneling but I haven't bothered to do that yet.

why not just use IPv6?

Someone will probably roll something like this out in the metro areas and it will eat the NBN

Webpass only serves multiple dwelling units AFAIK.

You can already order microwave and fixed fiber links in Sydney CBD, Pyrmont, etc but they cost a lot more than $60/m.

The NBN was a solid idea. Execution is always kind of the issue though.

> was

:( such a sad truth.

I doubt it. I'm no expert, but isn't it illegal to provide non-NBN internet to any house that has NBN?

I recall someone complaining that they had no internet at all for months, because the NBN wire got hooked up to their house, but not plugged into anything at the other end - then their internet company was legally obliged to disconnect their DSL because technically they had an NBN cable - it just wasn't plugged in to anything.

Of course this would be a good law IF we were getting proper fibre NBN (and the above technicality got seen to). Just like we don't want multiple companies running power lines to one house.

It's not illegal for a company to offer an alternative to the NBN. I think it is however mandatory for a provider who can service a dwelling with NBN to upgrade from a legacy technology (i.e. DSL) to the NBN.

This does not preclude other high-speed internet options servicing those dwellings.

For example, https://www.lightningbroadband.com.au offers or plans to offer 100/100 (actually, some pages have mentioned 200/200 but maybe that's coming later or business only). A salesperson explicitly told me they plan to service NBN-connected areas too but that they'll probably prioritise other non-NBN metro areas first, because the demand will be stronger.

Based on what I know of Webpass' network architecture, I'd be very surprised if latency was >5ms average from any of their individual on-net buildings back to any of the major IX points where they have routers in the SF Bay area.

Former Webpass customer here. I never used it for gaming, but the highest latency I ever got from SpeedTest with an SF server was 8ms and it was usually 3-4. My building was on the 100mbps plan and the actuals were ~90mbps down, ~65mbps up.

I'd hope that connection would be good enough for gaming.

after like 100kbps speed does not matter for gaming. you need a good quality connection (stable latency + no packet loss)


It's point to point wireless to a building and Ethernet from there. No real latency issue there.

Curious what they're using for backhaul wireless links if they're offering 1 gbps service. Not many solutions can handle that kind of speed.

Right now you can get away with a lot of "oversubscription" in MDU situations when you have 1-2Gbps to the building (typical millimeter-wave radio speeds).

I wouldn't be surprised if they have buildings fed with 1x1Gbps going to 40x1Gbps clients.

(hand waving about Erlangs and queuing theory here) E.g. if you assume everyone is downloading a DVD image from a fast mirror (CentOS install DVD or something), they probably won't all start the download at the same time so there's a really good chance one user will finish their download before someone else starts theirs.

My MDU users rarely crack 1Mbps on a 30-day 95th percentile basis.

Are you preparing for popular streaming video products pushing 4K content?

A popular model is to use wireless until enough you have enough customers in a building to justify building out fiber so you can supply more bandwidth. Even without fiber you could have a 50 unit building with half the units streaming 4K@30Mbps simultaneously and have enough bandwidth left over for typical web use.

Looks like millimeter wave: https://webpass.net/about_network.

Weird, I had Webpass for a year and was using it for a ton of gaming, and had no idea it was using wireless. It was really consistent and fast.

I've got about 15ms to google.com (and the silly speedtest.net page reports 4ms for whatever they chose nearby).

The latency to the West Coast Battle.net server is a very steady 12ms according to the Starcraft in-game UI.

I live in SF and have the 100mbps Webpass. Over my wifi, I get pings of <5ms to AWS us-west-1.

Webpass has only been doing fiber for 10 months [0], as far as I know the bulk of their customers in SF are on fixed wireless. I'm discouraged by this announcement - now it seems there's zero chance of widespread Google Fiber in SF.

[0] https://webpass.net/blog/fiber-is-here

I'm surprised by your downvotes. While "zero chance" might be a bit dramatic, it seems like this warrants discussion. Does this mean that Google thinks P2P wireless is the way to cover urban areas? Will it disincentivize them from persevering through a city-wide fiber installation because P2P wireless is "good enough"? Or is it a good opening play for fiber in SF because Webpass was already navigating the zoning/regulation/rights and just needed capital to succeed? I'd be very interested to hear others' perspectives on this.

Disclosure edit: I'm a current Webpass P2P wireless customer. I was literally about to switch away from Webpass to another P2P wireless solution (we have a site evaluation next Monday) because Webpass has been unreliable for us lately. I'm really curious what, if anything, this means for Webpass and Google Fiber in SF.

I don't really get this phobia of fixed wireless. The technology is improving by leaps and bounds--while the technology for digging up city streets isn't.

The city hasn't been doing a great job of dropping in fiber either. Maybe in SoMA, but up here in the hills I had my street repaved and there was no plans to put in fiber or upgrade sewer while it was being dug up (I checked with DPW). Fixed wireless is the future in SF, even with SF's dig once policy (http://sfgov.org/dt/dig-once).

I'd argue that the technology for digging up city streets hasn't needed to improve (though the politics around zoning/regulation/rights probably could); whereas, the technology for fixed wireless has--by leaps and bounds. But I argue this lightly because it detracts from my original question: Is fixed wireless the future of Google Fiber in SF? I'm not presupposing an answer, nor am I equipped to make a value judgement on said answer. Just curious & felt my parents' question deserved discussion, not downvotes.

The resources needed to widescale deploy fixed wireless are an order of magnitude less than residential fiber. I was counting on Google flexing its muscle to accomplish the fiber that less powerful organizations could not push through.

Google doesn't have the muscle to deploy fiber in SF on terms that make economic sense. If Fiber is profitable, a big part of the business case undoubtedly depends on the conditions Google imposes as a prerequisite for deployment: fast-track permitting and no build-out requirements. It'll take a lot of leverage to get SF to agree to those things.

At the same time it's kind of a huge embarrassment for Google not to offer Fiber in its own back yard. Fixed wireless gives Google a bridge to offering Google Fiber service in SF without taking on the SF government head on.

That's not entirely correct. Webpass has been using glass provided by other companies for years, that's how they get the redundancy in the mesh.

In 10 months they've been laying down their own glass trunk lines along a VERY small corridor. Last year was ~15 blocks of Howard St. This year was supposed to be Bryant St. So if you were in SOMA you had a decent chance.

Google has already stated they were going to use some other providers glass along with deploying their own in the city. This seems like a really good way to jump start things for GF in SF.

(I'm not an industry expert) Prior to 10 months ago, were there any Webpass end users without the fixed wireless hop to fiber? What exactly does Webpass mean when they distinguish between their fixed wireless product and their fiber product?

I'm not concerned with who owns the fiber, I would just prefer not to have wireless involved because of weather and oversubscription.

"were there end users on fiber?" Yes. Not many but yes. A large amount of their wireless backhauls terminated at data centers in SF, but some also "terminated" at buildings. These locations you could use webpass and basically be on fiber direct - though that wasn't the purpose. It was mostly so that they didn't have to backhaul connects ALL the way across the city and should a DC or major link go offline, it could reroute. Yay mesh networking.

The difference between that and fiber is who owned the glass. All of what I mentioned above, to my knowledge, is glass leased from other companies (Ex: AT&T). Their fiber product is glass they dug up the street and laid themselves - directly from the DC. If you look at the fiber maps in SF ( https://webpass.net/fiber ) they all start at 365 Main ( http://365datacenters.com/ ).

As a webpass customer, I have to say that weather and oversub have not been an issue for me for the last 3 years. For example this winter there was only 2 days where it rained hard enough (when it was REALLY coming down) to notice latency. In both cases it was for ~10mn and I only noticed because I was on SSH connections. Any muggle's netflix connection wouldn't even flinch.

I had Webpass fiber in Chicago for over a year. It was nice 1Gbps symmetric too. Back on Comcast now though :(

For those interested, you can get an idea of Webpass' network (with downloadable KMZ) by searching for their 70/80 GHz link registrations here: http://www.micronetcommunications.com/LinkRegistration/Query...

Search on Licensee Name Webpass

I'm a Webpass customer. Pretty amazing service. Although lately they have been suffering from some slow speeds at peak times. I think its just growing pains due to large increases in their customer base. I believe it will be back to high speed soon.

speedtests put me at around 100Mbps consistently when its "fast" and about 20-30 Mbps when its slow.

I hope it's not a sign that Google is slowing down fiber rollout and plans to switch to "new hot" wireless, but rather they'll push both.

does anyone know the eta for los angeles?

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact