Friday, June 30, 2006
Thursday, June 29, 2006
Wednesday, June 28, 2006
7.9 x 10^28
Garett Rogers had this blog about "Google's secret IPv6 plans". It appears that Google owns a block of IPv6 addresses numbering approximately 7.9 x 10^28 (79 billion billion billion addresses) or 2^96. Basically Google owns any IPv6 address from:
2001:4860:0000:0000:0000:0000:0000:0000
to
2001:4860:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF
ARIN (the organization that allocates IP addresses on the Internet) only gives IP blocks of this size to ISP for the purpose of reselling to end users and companies. Put this in the context of Google's continued purchase of dark fiber (unused fiber optic cabling, much of it left from the dotcom era) and the construction of a massive data center, it's clear that Google is trying to position itself in the heart of the Internet.
2001:4860:0000:0000:0000:0000:0000:0000
to
2001:4860:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF
ARIN (the organization that allocates IP addresses on the Internet) only gives IP blocks of this size to ISP for the purpose of reselling to end users and companies. Put this in the context of Google's continued purchase of dark fiber (unused fiber optic cabling, much of it left from the dotcom era) and the construction of a massive data center, it's clear that Google is trying to position itself in the heart of the Internet.
Hierarchy among geeks, nerds, and dweebs.
This is taken from comments on a Slashdot article talking about the Best Buy "Geek Squad". This guy is definitely jealous of engineers:
"What follows, I claim, is the one true classification of geekdom. It has stood up to rigorous peer review (loud arguments amongst drunken physics students) for years, and I stand by it.
A dweeb is someone without social skills who either doesn't recognize or is unable to accept that they are unusual. They constantly *try* to fit in, with disastrous results, and dedicate a significant portion of their daily lives to obsessing over how to pass as normal.
A nerd is someone without social skills or popular interests who recognizes that he or she is unlike most people and feels no shame in it.
A geek is a nerd with technical skills and passionate interests; in particular one who has a myopic dedication to a particular specialty. (This is the subspecies *true geek,* distinct from but related to the *common geek,* or nerd who is generally technically savvy and useful to have around.)
To summarize, the dweeb is the guy wearing a slightly out of fashion hipster shirt who generally creates embarrassing silences at parties by saying awkward things about pop stars or sports teams.
The nerd is the guy who skips the party in order to achieve moderately high scores on a popular video game while eating unheated canned peas with a spoon and listening to recordings of experimental music.
The geek is the guy who skips the party in order to code a popular video game, figure out the angle of repose one might expect for a pile of canned peas, or compose and record some experimental music.
On the college campus, geeks make up virtually the entire population of physics and math majors (as well as a majority in classics, many of the less trendy engineering sub-disciplines, linguistics, physical anthropology, and some of the more obscure languages.)
The nerds are the guys who drop out of school after one semester but stay in a college town working in a bookstore, where they get great discounts on whatever genre books they happen to like and talk to their geek friends about writing their own books yet never seem to actually finish any of them.
The dweebs largely end up in engineering or the quantitative business disciplines, in the hopes that they can earn enough money to buy the respect of powerful and attractive people. Those in engineering have a tough time of it, as they are publicly ignored by the normals whom they so admire while simultaneously earning the scorn and contempt of the geeks in their departments. Those in business do rather well, since they have a good chance at fooling their colleagues into thinking that they are geeks. (Normals may not invite geeks to parties, but they do like to hire them.)"
Tuesday, June 27, 2006
Kraftwerk's vocoder on eBay
A prototype vocoder built for Kraftwerk and pictured on the album Ralf & Florian is up for auction on eBay. Starting bid is $3,800. From the auction description:
Pictures 1The device was used for two studio productions, title "Ananas Symphonie" including synthetic vocals and rhythm track in conjunction with lapsteel guitar and rhythm machine , as well for "Kristallo" with rhythm machine and EMS-Synth.
Later used to sound the KW's intro for "Autobahn..."
It was conceived , built and constructed by electronic engineers in the area of Physikalisch-Technische BundesAnstalt (PTB) Braunschweig, Dipl.Ing P.Leunig and Dipl.Ing. K.Obermayer.
Later the well known german music studio engineering and supplier company R.Barth K.G.,Hamburg, Germany, boiled down the know-how gained from this project into so called MUSICODER which was produced in small quantities.
Pictures 1The device was used for two studio productions, title "Ananas Symphonie" including synthetic vocals and rhythm track in conjunction with lapsteel guitar and rhythm machine , as well for "Kristallo" with rhythm machine and EMS-Synth.
Later used to sound the KW's intro for "Autobahn..."
It was conceived , built and constructed by electronic engineers in the area of Physikalisch-Technische BundesAnstalt (PTB) Braunschweig, Dipl.Ing P.Leunig and Dipl.Ing. K.Obermayer.
Later the well known german music studio engineering and supplier company R.Barth K.G.,Hamburg, Germany, boiled down the know-how gained from this project into so called MUSICODER which was produced in small quantities.
Monday, June 26, 2006
Neutracide: Net Neutrality: Part XIII
By Cory Doctorow, InformationWeek
June 26, 2006
What Is Neutral, Anyway?
Additionally, the rules are going to have to do three incredibly tricky things:
1. Define network neutrality. This is harder than it sounds. If a Bell lets Akamai put one of its mirror servers in a central office, then Akamai's customers can get a better quality of service to the Bell's customers than those using an Akamai competitor. This is arguably a violation of net neutrality, but how do you solve it? It's probably not practical to require the Bells to let all comers put local caches on their premises; there's only so much rack space, after all.
Another tricky case: the University that provides a DSL service to its near-to-campus housing and configures its network to deliver guaranteed throughput to a courseware archive. It gets even stickier if the DSL and/or the courseware archive are supplied by commercial third parties. Poorly written net neutrality regulations could prevent universities from providing those services, which should be allowed.
This is just the tip of the iceberg. Imagine a Bell that came up with a plan to let services tune their applications to improve on throughput to its customers. For example, some Bells might be able to tune service more efficiently by providing real-time feedback to companies about the optimal frame size for its network, or information on traversing its internal private networks, and so on. This is a good way to wring performance out of lines and switches, but the efficiencies decay if the Bell is legally required to provide this service to every customer, without being compensated for the additional effort.
2. Revise the definition of network neutrality in a way that keeps pace with new network realities. Next week, someone will come up with a way of tweaking the Internet that may not be neutral. It will happen again the week after, and the week after that. Whichever regulator gets to play neutrality czar will have to make rules that keep pace with those changes.
The definition must avoid the risk that the rules will fail to keep up with advancing technology, allowing the creation of a tiered Internet by means they today's regulators don't envision.
Even worse is the risk that the Bells (or someone else, like a far-future, calcified and thoroughly evil-fied Google) will get the rules changed to keep new entrants out of the market by declaring their businesses non-neutral.
3. Most difficult of all: Know when nonneutrality is committed, and stop it. It's not enough to prohibit the Bells from advertising the rate for tiered access to their pipes. The Bells have a long, dishonorable history of backroom deals with spammers, the National Security Agency, and other unsavory sorts.
June 26, 2006
What Is Neutral, Anyway?
Additionally, the rules are going to have to do three incredibly tricky things:
1. Define network neutrality. This is harder than it sounds. If a Bell lets Akamai put one of its mirror servers in a central office, then Akamai's customers can get a better quality of service to the Bell's customers than those using an Akamai competitor. This is arguably a violation of net neutrality, but how do you solve it? It's probably not practical to require the Bells to let all comers put local caches on their premises; there's only so much rack space, after all.
Another tricky case: the University that provides a DSL service to its near-to-campus housing and configures its network to deliver guaranteed throughput to a courseware archive. It gets even stickier if the DSL and/or the courseware archive are supplied by commercial third parties. Poorly written net neutrality regulations could prevent universities from providing those services, which should be allowed.
This is just the tip of the iceberg. Imagine a Bell that came up with a plan to let services tune their applications to improve on throughput to its customers. For example, some Bells might be able to tune service more efficiently by providing real-time feedback to companies about the optimal frame size for its network, or information on traversing its internal private networks, and so on. This is a good way to wring performance out of lines and switches, but the efficiencies decay if the Bell is legally required to provide this service to every customer, without being compensated for the additional effort.
2. Revise the definition of network neutrality in a way that keeps pace with new network realities. Next week, someone will come up with a way of tweaking the Internet that may not be neutral. It will happen again the week after, and the week after that. Whichever regulator gets to play neutrality czar will have to make rules that keep pace with those changes.
The definition must avoid the risk that the rules will fail to keep up with advancing technology, allowing the creation of a tiered Internet by means they today's regulators don't envision.
Even worse is the risk that the Bells (or someone else, like a far-future, calcified and thoroughly evil-fied Google) will get the rules changed to keep new entrants out of the market by declaring their businesses non-neutral.
3. Most difficult of all: Know when nonneutrality is committed, and stop it. It's not enough to prohibit the Bells from advertising the rate for tiered access to their pipes. The Bells have a long, dishonorable history of backroom deals with spammers, the National Security Agency, and other unsavory sorts.
Tim Berners-Lee on Net Neutrality: "This is serious." Net Neutrality: Part XII
The inventor of the WWW has a short, to-the-point post that explains exactly why supporting real, bona fide net neutrality is the Right Thing to Do. I absolutely encourage you to read the entire post, but really he sums up the whole argument for net neutrality in his opening sentence:
When I invented the Web, I didn't have to ask anyone's permission.
Ma Bell may I?
What the opponents of net neutrality are pouring millions into lobbying for is a world where, when someone offers a new high-bandwidth service over the Internet, they have to go around to each of the last-mile providers and ask, "may I have permission to compete on a level playing field with the other services that go over your pipes?" And if entrepeneurs can't come up with enough funding to appease the troll that guards that particular bridge, then they could effectively lose access to the customers at the other end.
To move the discussion away from the typical example of high-bandwidth video, let me turn to the example of a new email attachment service, called Pando, that was recently brought to my attention. From what I can tell, Pando is a P2P service for sending very large files through email. You attach to your email a .pando file, which is probably something like a .torrent file, and a user on the other end who has the Pando client can open that .pando file and use it to begin downloading the larger file that you want to send them.
Now, I have no idea how well this Pando thing works, or even if it works at all. I bring it up only because it seems like new, kind of slick start-up idea. But on a non-neutral Internet, Pando is a much less attractive as a small business. Why? Because if I'm on Comcast and you're on AT&T, and I try to send you a .pando file, then both of our ISPs have to have agreed to let this new service's packets through at a reasonable clip. If either Comcast or AT&T has decided to throttle Pando packets (maybe Pando is being used to send video files!), then Pando will work for me when emailing some folks but not others. In the face of such uneven and generally unpredictable results, most users will probably skip the service entirely, and the world would have one less small business.
At this point, I should insert a word on "throttling" vs. quality of service (QoS). The telcos are arguing that they don't want to throttle anyone's traffic—they just want the freedom to offer better transport (QoS) to services that pay up. This is a misleading way to frame the issue.
If three companies run competing video-on-demand services and only one has paid the piper, then the other two guys are effectively throttled as far as end users are concerned. This is especially true if the baseline level of transport service for streaming media on that ISP's network is set painfully low. All Comcast would have to do is make a rule that all streaming video sites default to a relatively small amount of bandwidth, and they get more acceptable levels of bandwidth as they pay more. This amounts to throttling people who don't pay, and I think the odds are greatly in favor of this kind of thing becoming rampant on a non-neutral Internet.
Even worse would be a situation where Comcast enters into some kind of exclusive deal with an online service provider, with the result that the cost of competing on a level playing field with that favored service are set prohibitively high. Of course, such a move would probably spark lawsuits from competitors, but do we really want to rely on very expensive tort law to enforce what simple regulation could do much more cheaply and rationally?
When I invented the Web, I didn't have to ask anyone's permission.
Ma Bell may I?
What the opponents of net neutrality are pouring millions into lobbying for is a world where, when someone offers a new high-bandwidth service over the Internet, they have to go around to each of the last-mile providers and ask, "may I have permission to compete on a level playing field with the other services that go over your pipes?" And if entrepeneurs can't come up with enough funding to appease the troll that guards that particular bridge, then they could effectively lose access to the customers at the other end.
To move the discussion away from the typical example of high-bandwidth video, let me turn to the example of a new email attachment service, called Pando, that was recently brought to my attention. From what I can tell, Pando is a P2P service for sending very large files through email. You attach to your email a .pando file, which is probably something like a .torrent file, and a user on the other end who has the Pando client can open that .pando file and use it to begin downloading the larger file that you want to send them.
Now, I have no idea how well this Pando thing works, or even if it works at all. I bring it up only because it seems like new, kind of slick start-up idea. But on a non-neutral Internet, Pando is a much less attractive as a small business. Why? Because if I'm on Comcast and you're on AT&T, and I try to send you a .pando file, then both of our ISPs have to have agreed to let this new service's packets through at a reasonable clip. If either Comcast or AT&T has decided to throttle Pando packets (maybe Pando is being used to send video files!), then Pando will work for me when emailing some folks but not others. In the face of such uneven and generally unpredictable results, most users will probably skip the service entirely, and the world would have one less small business.
At this point, I should insert a word on "throttling" vs. quality of service (QoS). The telcos are arguing that they don't want to throttle anyone's traffic—they just want the freedom to offer better transport (QoS) to services that pay up. This is a misleading way to frame the issue.
If three companies run competing video-on-demand services and only one has paid the piper, then the other two guys are effectively throttled as far as end users are concerned. This is especially true if the baseline level of transport service for streaming media on that ISP's network is set painfully low. All Comcast would have to do is make a rule that all streaming video sites default to a relatively small amount of bandwidth, and they get more acceptable levels of bandwidth as they pay more. This amounts to throttling people who don't pay, and I think the odds are greatly in favor of this kind of thing becoming rampant on a non-neutral Internet.
Even worse would be a situation where Comcast enters into some kind of exclusive deal with an online service provider, with the result that the cost of competing on a level playing field with that favored service are set prohibitively high. Of course, such a move would probably spark lawsuits from competitors, but do we really want to rely on very expensive tort law to enforce what simple regulation could do much more cheaply and rationally?
More accurate storm prediction: Geostationary Operational Environmental Satellites (GOES-N)
"Designed and built at Boeing’s Satellite Development Center in El Segundo, Calif. for NASA and NOAA, the GOES-N series spacecraft are based on the popular three-axis Boeing 601 model satellite. Technology on the GOES-N spacecraft will improve image accuracy by a factor of four using a more stable instrument platform and a more precise control system based on a geosynchronous star sensor attitude determination and control system know as a “star tracker.”
The capabilities aboard the GOES-N satellite will provide more accurate prediction and tracking of severe storms and other weather phenomena, resulting in earlier and more precise warnings to the public. GOES-N will support NOAA and NASA scientists collecting and analyzing real-time environmental data, as well as rescuers responding to calls for help with the assistance of a communication subsystem that includes search and rescue capability to detect distress signals from land, sea and air.
Boeing’s four-decade knowledge and experience in weather and Earth-observation space systems underpins the next-generation environmental system in support of the NOAA’s strategic mission of assessing and predicting environmental changes, protecting life and property, and providing decision-makers with reliable climate information."
The capabilities aboard the GOES-N satellite will provide more accurate prediction and tracking of severe storms and other weather phenomena, resulting in earlier and more precise warnings to the public. GOES-N will support NOAA and NASA scientists collecting and analyzing real-time environmental data, as well as rescuers responding to calls for help with the assistance of a communication subsystem that includes search and rescue capability to detect distress signals from land, sea and air.
Boeing’s four-decade knowledge and experience in weather and Earth-observation space systems underpins the next-generation environmental system in support of the NOAA’s strategic mission of assessing and predicting environmental changes, protecting life and property, and providing decision-makers with reliable climate information."