Forgot your password?

Comment: Re:It's not arrogant, it's correct. (Score 5, Insightful) 466

by Burdell (#46567605) Attached to: AT&T Exec Calls Netflix "Arrogant" For Expecting Net Neutrality

Netflix pays for their bandwidth

Well, but they don't always, at least not as much as anybody else. Several times in recent years, Netflix has switched bandwidth providers to "wanna-be tier 1" networks; that is, networks that are not as well-connected as they'd like to be because they don't really meet anybody's requirements for settlement-free peering. These providers see Netflix as leverage against their bigger competitors and appear to have sold Netflix bandwidth at well market prices in order to strong-arm competitors to provide new network interconnects.

Large networks don't just peer with anybody. There are costs involved in each additional turn-up, both for hardware ports and for the management side. They also don't just peer at a single or few locations (since that can allow outsider actors to cause drastic changes in internal network bandwidth utilization); they require other large networks to peer in a bunch of different places. Some of the smaller networks can't afford to do that, and want to dump large traffic hogs like Netflix at already congested peering points, and then complain that the big guys didn't bend over backwards to help them.

I've worked for small to very-small ISPs for over 18 years, and I definately don't hold Netflix blameless in this. They do things they know will impact their customers and then blame the other networks for all problems (and they aren't the only one, just one of the biggest in recent years).

Comment: Re:Possible botnet C&C related (Score 2) 349

by Burdell (#46458751) Attached to: Crowdsourcing Confirms: Websites Inaccessible on Comcast

CNAME on the root record of a zone is not allowed. .org servers delegate to ns1/ with NS records, so ns1/ must supply an SOA and one or more NS records for Instead they provide an out-of-scope SOA, valid-looking A, MX, and CNAME (which is also a bogus combination) but return NXDOMAIN for NS.

The real answer is that ns1/ have a wildcard for * with A, MX, and CNAME records. Somehow they also respond to any SOA request with an SOA for, and have no NS records.

I still suspect a botnet C&C DNS server is running, with probably a rapidly-changing set of domains delegated to it. Comcast is probably blocking delegations to those servers, and the only real choice (that isn't a lie) for DNS responses would be SERVFAIL (in this case due to policy). NOERROR+no ANSWER records or NXDOMAIN would not really be true.

Comment: Possible botnet C&C related (Score 3, Informative) 349

by Burdell (#46457025) Attached to: Crowdsourcing Confirms: Websites Inaccessible on Comcast

The DNS for is rather fishy looking. The .org servers have NS records pointing to and, which have a 20 second time to live (vs. a normal 1 day TTL), which is common in botnet command & control networks. Also, the ns1/ servers give answers to A lookups, but return NXDOMAIN for NS lookups (which is completely bogus; NXDOMAIN means that does not exist, not that it doesn't have NS records, which would still be bogus).

The NXDOMAIN for NS records would cause many caching servers to cache NXDOMAIN for all records (not just NS), which would cause the domain to not resolve (depending on the order things were looked up). Basically, I don't see this as a Comcast problem, but rather a problem with the DNS servers for This may be accidental (although AFAIK no normal DNS server would reply with A records but return NXDOMAIN for NS records), but looks possibly like it is intentional and possibly part of a botnet C&C. There's a lot of that going on lately.

Comment: Re:Other options? (Score 1) 247

by Burdell (#46364173) Attached to: The Rescue Plan That Could Have Saved Space Shuttle <em>Columbia</em>

Oops, yeah, I forgot Apollo 7. They probably would have been able to survive, although it might have been rough. The biggest problem probably would have been that they would not have had much choice in where they landed (could have ended up in a location where recovery was effectively impossible or would take too long, could have hit land instead of water, etc.).

Comment: Re:Other options? (Score 1) 247

by Burdell (#46352489) Attached to: The Rescue Plan That Could Have Saved Space Shuttle <em>Columbia</em>

There are risks in spaceflight that just can't really be overcome, except in hindsight. If what happened to Apollo 13 had happened to Apollo 8, the result would have been very different. Apollo 8 had no LM that could have been used as a "lifeboat", and it is unlikely that there would have been any other way to keep the astronauts alive. There's a good chance the Apollo program would have ended if NASA had two consecutive crews killed.

However, one thing from Apollo 8 helped Apollo 13: on Apollo 8, Jim Lovell accidentally erased the flight computer's memory and had to re-figure the position from start sightings. He had to do a similar task during Apollo 13 after the computer was powered down and restarted.

Comment: Re:I don't understand (Score 1) 363

by Burdell (#45442907) Attached to: Arizona Approves Grid-Connection Fees For Solar Rooftops

Let's say you and I can both buy a shelf at Wal-Mart for $10. Now I start making shelves for myself instead, and make an exact duplicate of Wal-Mart's $10 shelf. Should my nearest Wal-Mart be required to buy my shelf for $10, transport it to your nearest Wal-Mart, and then sell it to you for $10? They have trucks already, so why should they charge me for the transportation costs?

I /suppose/ that AP might be operating at loss here if they have to pay out more per watt than it costs them to generate it themselves

That's exactly the case. If they charge residential customers $0.10/kWh, you don't think all $0.10 goes to pay for the power plant, do you? They have to transport the power from the plant to the customer's location (which has loss in the system; they have to generate more than 1 kWh to deliver 1 kWh), they have to meter how much the customers use, bill for the usage, maintain the system, etc.

Pick-up and deliver only makes sense when you get more for the delivery than you pay for the pick-up.

Comment: Re:I think it's more likely a Cogent problem. (Score 3, Insightful) 202

by Burdell (#44045355) Attached to: Verizon Accused of Intentionally Slowing Netflix Video Streaming

Verizon doesn't "choose" ISPs; they _are_ a backbone provider (they don't buy transit from anybody). Cogent is known for peering disputes, as well as selling hard to content providers (and sometimes eyeball networks) they think will give them leverage in peering disputes.

Smaller ISPs (that do buy transit) know that you don't buy from Cogent unless you have at least two other paths to everything on the Internet.

Comment: Re:More pertinent information on beer fridge (Score 1) 231

by Burdell (#43902043) Attached to: Beer Fridge Caught Interfering With Cellular Network

I don't know about Australia, but in the US, you are responsible if you are causing interference in somebody else's licensed band. Even if you didn't mean to, you are transmitting (noise) on a licensed frequency without a license. If it even looks like it might be because you made some modifications to radio gear, you can be liable for a large fine (and depending on the band possibly jail time).

I remember a few years ago a convenience store near Miami's airport was closed by the FCC because they had some dodgy electronics (door opener or bar-code scanner) that was interfering with the air traffic control radio frequency. The FCC forced the store to close (and IIRC shut off the power because they weren't sure of the source device) until they could prove they had non-interfering equipment.

Even the ISM (the so-called "unlicensed") bands, like 900 MHz and 2.4 GHz, are very tightly controlled. Equipment must stay within regulations on channels, power limits, and usage. Your equipment must be able to handle interference from other devices operating within the regulations, but if somebody operates outside the limits, they are liable and must shut down (and face fines, etc.).

Comment: Re:Getting to 24-48 hr advance warning (Score 5, Informative) 104

by Burdell (#42931999) Attached to: ATLAS Meteor Tracking System Gets $5M NASA Funding

For something like this (where nobody died), you wouldn't attempt an evacuation. I believe that most of the injuries were from broken glass and other falling debris; it would be enough to warn people to either get outside (away from buildings, trees, and other objects that could be blown around by a shock wave) or to stay inside away from windows.

Comment: Re:How fractured is ARM? (Score 4, Informative) 151

by Burdell (#42251809) Attached to: Linux 3.7 Released

There are variants in the instruction set (just like there are in the x86 world, where i686 is a superset of i383 for example). However, that isn't the big problem with ARM; there isn't a single-standard way of booting like there is with x86 (where most things are IBM PC BIOS compatible, with some now moving to EFI/UEFI). Also, there's no device enumeration like ACPI; lots of ARM vendors build their own kernel with a static compiled-in list of devices, rather than having an easy way to probe the hardware at run-time.

"Never face facts; if you do, you'll never get up in the morning." -- Marlo Thomas