Verizon Wireless announced Friday that it was paying $3.6 billion to three cable TV companies — Comcast, Time Warner Cable, and Bright House Networks — in exchange for wireless licenses the companies bought in an FCC auction in 2005. Pundits are describing the deal, and especially its cross-marketing provisions, as revolutionary with the potential to change the way we communicate and are entertained. I doubt this. Rather, I think it reflects a failure of the cable companies to compete in other markets.

I remember this license auction and wrote about it at the time. New spectrum was being released and the MSOs were afraid Verizon and AT&T would snap it up to compete with them for video. So the cable companies bought the spectrum specifically to keep it out of commerce, which has been the case now for six years.  Mission accomplished. Yet the price they are getting from Verizon when you wrestle through the apples and oranges of varying licenses bought and sold at different times isn’t significantly higher than they paid back in 2005, especially since the licenses have produced zero cash flow for going on seven years.

But the point isn’t the price, we’re told, but the co-marketing — that Comcast can sell Verizon Wireless service eventually without even calling it Verizon or that Verizon will presumably resell Time Warner Cable or Bright House to the very folks it would rather buy Verizon’s own FiOS video package.

Yeah, right.

I think the co-marketing story is just spin and that not much will come of it.

We’re heading into a bandwidth war between DOCSIS 3 cable modems and fiber-to-the-curb services like FiOS. And while I can’t predict which side will win or lose this battle I can say that wireless service won’t be a factor in the decision.

Yes, we get e-mail and play Angry Birds on our smart phones and yes, 18 months from now all mobile phones will be smart phones, but the mobile transition hasn’t had any impact on our bandwidth use in homes or offices. It has just given us a way to consume even more electrons at lunch or in the car.  And for all the very real potential of Long Term Evolution (LTE) 4G networks, they can’t in practical terms serve enough bits to enough people at the same time in any city to be viable competition to almost any form of wired Internet, whether from the phone company or the cable company.  The physics just doesn’t support it.

Eighteen months from now many American homes will be where Japanese and Korean homes have been for sometime, sucking 100 megabits or more from the Internet. LTE can’t do that now or then and it can’t do half of that for a tenth or even a hundredth of the customer base of wired Internet.

Verizon needs the bandwidth because voice landlines are going away and it has to compete with AT&T, not Comcast. The new voice is all wireless. But at the same time, even 4G wireless will come to share analog voice’s sense of not being enough.

FiOS was deliberately designed from the very beginning with the good glass — fiber that can go to a gigabit and beyond. DOCSIS 3’s channel bonding and network segmentation will eventually allow full access to 70+ video channels for data service where currently most cable modems use one channel and some use two or three. No matter what way the wire gets to your house that wire will soon carry 100 megabits and then a gigabit that LTE never will.

Those wireless bandwidth caps are there for a reason.

Cable companies today make most of their profit from providing Internet service. FiOS and similar services are Internet services enhanced to keep subscribers subscribing, not especially to give them Wheel of Fortune.

This deal is the cable companies getting out of wireless because they can’t figure how to make money in that business.

Podcast: Play in new window | Download

There’s a dispute going on right now between Comcast and Level3 Communications concerning the peering agreement between those two companies. Comcast says the dispute has nothing to do with the fact that Level3 just got the Netflix video streaming contract while most observers think that’s all it has to do with.

I think so, too.

Peering is at heart nothing but restraint of trade. Peering came about when various Internet backbone providers noticed they were all connected to the same big data centers and points of interconnection, normally inside telco central offices. Simply pulling an Ethernet cable from one rack to another could interconnect millions of users from two different backbone providers, saving time, distance, router hops and total bits in the process. Peering agreements typically involve no exchange of money since they are intended to be between peers — very similar companies of roughly comparable size that would be sharing equal numbers of bits back and forth. Peering agreements were for big companies, especially backbone providers interconnecting with the fundamental idea that they’d be giving as many bits and they got and therefore no direct compensation would be required. It also kept smaller companies out of the backbone business because they were made to pay, and dearly.

Level3 is mainly a backbone company that is lately delivering a lot of streaming video, too. Comcast points to the disparity between the number of Netflix video bits served (a lot) to the number received (almost none for Netflix other than some Quality of Service data and of course the movie orders). That’s not the deal, says Comcast, which wants Level3 to pay the difference in cash.

On the other hand, Comcast for the most part isn’t an Internet backbone provider. They have some backbone assets, sure, but mainly they are America’s largest broadband ISP. So while Comcast can fault Level3 for taking advantage of their peering agreement terms, Level3 could as easily drop peering with Comcast altogether, still getting to Comcast viewers through other peers, though with the addition of some latency from the extra hops required.

Note that Netflix formerly did its streaming through Akamai’s Content Distribution Network (CDN) which shares revenue with participating ISPs.  Level3 probably got the Netflix gig by beating Akamai on price and they beat Akamai on price because they are relying on that darned peering agreement to make it possible.

As an ISP, Comcast could afford to drop one backbone, but not all of them, so Level3 has some power here — more than many commentators have noticed.

There is a lot of posturing here, so let’s try to figure out the real issue, which I think is Google.

Google has long wanted to drop a rack or a container or at least its own fiber connection two hops from every broadband user in Ameica and eventually the world. They’d like to do that through peering agreements like Level3 and certainly have as much of an argument as Level3 has for doing so, given Google’s own fiber assets, which are certainly more than Comcast’s. But Google will pay for access if it must, because global domination is worth the price.  The search giant is willing to pay if it must for guaranteed access.

Comcast knows this. As America’s largest broadband ISP, Comcast stands to gain more than any other company from allowing Google to run fiber into every head-end data center the company has. But Google won’t pay if they don’t have to. So to make sure Google pays, Comcast has to make sure Level3 pays.

That’s all it is. Both sides are distorting the peering agreement like crazy to make their points, which aren’t about equity, net neutrality, user rights, legal theory, who is actually paying for the bandwidth (customers), or anything else — just Google’s money.

Podcast: Play in new window | Download

I’d been putting-off going to startups.cringely.com to finally read all 286 entries so far in this summer’s Cringely (NOT in Silicon Valley) Startup Tour.  But when I finally went to the site, I couldn’t get in.  The page timed-out.  This was not good.  Or maybe it was very good in that the site was so busy.  But even that’s not good because I don’t like turning readers away.  So which was it — good or not good?

Not good.

Twelve hours later, when I still couldn’t get in I called the CTO at the company that hosts that site — Democrasoft.  You haven’t heard about them, believe me, and I’ll explain why below.  But they weren’t having any trouble seeing the site.  Nor was I having trouble seeing it on my iPhone, or using my Verizon MiFi cellular access point.  It seemed to be a problem with my home ISP — Comcast.

Twenty minutes on the phone with Comcast tech support found the problem, though not the solution: the server IP address was blacklisted by an outfit called SORBS (Spam and Open Relay Blocking System) that claims to keep track of mail servers run by spammers or compromised by computer crackers.  Of 105 such blacklists available, only one — SORBS — listed this IP address, which wasn’t even for a mail server!

That IP was part of a block of addresses owned by Amazon Web Services, with the entire block listed by SORBS as suspect.  So Comcast (only the eastern half of Comcast, I later learned), tending to believe SORBS, blocked the address and Cringely’s (NOT in Silicon Valley) Startup Tour from tens of millions of subscribers.

The suspect IP address may have been used previously for another machine that was a mail server or maybe a compromised web server.  It has only been startups.cringely.com for a couple weeks, after all.

I’m amazed to learn that as rigorous an outfit as Amazon Web Services doesn’t check its IP addresses for blacklist status before reassigning them.  If I were an AWS customer I would be upset.  Since I’m freeloading I guess I’m just a little miffed.

We’ll sort this out shortly, I’m sure.  The guys at Democrasoft have lodged a protest with SORBS, but I am not very confident that will accomplish anything quickly.  Better to make Amazon assign the server a different IP address.

If you are having trouble reaching startups.cringely.com as a result, try it from a computer with a different ISP.

In the meantime, what is this Democrasoft?  Well until a moment ago it was called Burst.com, a little company from Santa Rosa, CA that I wrote about years ago over and over when they were fighting Microsoft and then Apple in court, winning both cases.  Burst was involved then in the efficient distribution over the Internet of video and audio streams and I suppose they continue to own and license patents in that area today.

A few months ago the folks at Burst called to tell me they were changing direction, creating a new kind of web service designed to help groups explore issues and make decisions.  The called it Collaborize.

Startups.cringely.com, if you can get to it, is a custom instance of Collaborize dedicated solely to the nomination, discussion, and evaluation of startup companies.  Not even a beta, I’d say my site was alpha software, but I was intrigued by the concept, trusted the people behind it, and who can turn down free service?

Collaborize was formally announced this week at the Demo conference in Palm Desert.  I wasn’t there but from what I hear the product was well received with attendees seeing all sorts of interesting ways to use it.  Have a look and let me know what you think.

That is if your ISP will let you.

Podcast: Play in new window | Download

This week the U.S. Federal Communications Commission (FCC) releases its proposed new rules for Internet Service Provider (ISP) network neutrality.  I have written many times about Network Neutrality and once I have a look at the FCC proposal I am sure I’ll have comments to make here.  In general I’m in favor of rules that allow me, as a consumer, more digital freedom. It would be great to run Skype over my iPhone, for example, just as I can already run it over the cellular connection on my notebook. But right now I’m talking about a different kind of network neutrality, the kind I’m struggling to achieve in my own home.

I live in Charleston, South Carolina where my primary ISP is Comcast. I have a 16 megabit-per-second (mbps) business Internet service with five static IPs and an upstream speed that I think is supposed to be 2.0 mbps but actually measures around 2.5. On the Speakeasy Speed Test I have no problem clocking the full 16 mbps to Atlanta, either.  It’s not Verizon’s FIOS, it costs three times as much as FIOS, but my connection more than does the job.  Compared to some other places in the world of course my speeds are laughable.

So why is it that when I surf the net while speaking on my Voice-over-IP (VOIP) telephone, it breaks up?  It’s not like I don’t have enough bandwidth, both up and down. And the network in my house is 100 mbps wired Ethernet using Cat5 cable throughout.  Ah, but I’m using the hated Vonage telephone service you say, not Comcast’s VOIP offering.  That explains it: net neutrality violation!!!

Except it isn’t. Comcast and Vonage have been pretending to be friends for a while now. It’s all part of the “We don’t really need that old Net Neutrality” song Comcast and the other big ISPs have been singing, including the verse that says Vonage is okay by them.

Then why does my Vonage-connected fax machine not function reliably, either?

Maybe I need traffic shaping, you say. Let’s just adjust my router to give priority to those VOIP packets, as I am sure Comcast would do if I were using their service.

Except I already do traffic shaping. I run a rather robust firewall as a sort of Internet gateway that includes local DNS and Squid (proxy) service. VOIP Packets get first dibs on my cable modem and always have.

This problem has been driving me crazy for some time now, but I believe I know what’s happening and it has nothing to do with Comcast or net neutrality.

I’m pretty sure the problem is in the Vonage boxes that connect my phone and fax machine to the network, called Analog Telephony Adapters or ATAs. First, I don’t use my ATA’s as Vonage suggests. Vonage envisions a single-ATA network generally with a single PC, or at least they did when I got these puppies. They want me to plug my ATA into the cable modem and my PC into the ATA so the ATA automatically takes precedence. I can’t do that for three reasons: 1) my office is three floors above my cable modem; 2) my fax machine is not in the same room as my PC, and; 3) I’m pretty sure the Vonage Ethernet ports are limited to 10 mbps so hooking-in there would limit the bandwidth available to my PC. If I’m paying for 16 mbps, dag nabbit I want to use 16 mbps!

Given that I’m already doing traffic shaping in the router and have a huge excess of bandwidth for VOIP anyway, what’s the big deal using the ATA’s as I do, simply plugged into a 10/100 Ethernet switch? It shouldn’t matter.

Then I spoke with my friend Paul and came to a sudden realization. I’ve been messing with my Internet gateway, trying to convert it to a trio of $99 SheevaPlug computers that I’ll run as a tiny cluster just to see if I can do it. Paul said his testing showed each 1.2 GHz Sheeva was the equivalent of about a 10^th of his four-core AMD box. “But even that’s plenty to saturate an Ethernet connection,” he said.

The Sheeva installation isn’t even ready to go yet, but what came to me is that the poor Vonage ATAs just can’t keep up. I got them when I signed up for Vonage service in 2002!  Back then my computer had a single core and ran at 400 MHz. Today I have four cores and run at 3.0 GHz. While it technically isn’t supposed to work that way I’m guessing my PC is just so darned fast at grabbing and releasing bandwidth those little seven year-old Motorola ATAs from Vonage are having trouble getting a packet in edgewise. Yes, the switch should compensate for that but you know I think that switch is about seven years old, too.

That explains why VOIP clients like Skype and Gizmo that run entirely on my PC (no ATA) don’t have any problems.

Most of my hardware is replaced every three years, but these network components have been running undisturbed since they were first installed. And being digital they probably run as well as ever. They just weren’t built with the idea that one day there would be a bully in the house.