2) They are "forcing" adwords customers to have their ads listed on "link sites"... No they aren't. You can opt out of AdWords (and indeed its usually a good idea).
No, it will not. Peering is based on BGP, which doesn't do any sort of dynamic load balancing. Its effectivly a link-state protocol, as long as the peer with the best path (and other metrics that're not related to the current performance of the link) is still up, the traffic will prefer that peer, regardless of throughput.
However, peering agreements typically have clauses that prevent any sort of traffic shaping.
"A hundred thousand harmless junkies became criminals overnight, by Act of Congress, in nineteen twenty-seven. Ten years later, in thirty-seven, all the pot-heads in the country became criminals overnight, by Act of Congress. And they really were criminals, when the papers were signed. The guns prove it. Walk away from those guns, waving a joint, and refuse to halt when they tell you. Their Imagination will become your Reality in a second."
-Illuminatus, Robert Anton Wilson
Specifically, encoded network traffic data collected in a packed form of unsigned 16, 32 and 64 bit integers. While its possible to shift the numbers to be signed, it makes the supporting code substantially more complicated, since ultimatly most of these numbers are not counters, they represent identifiers, which means all queries and output would need to do these calculations as well. Certainly not impossible, but its a huge increase in code complexity.
If postgres supported native unsigned integer types, I'd switch to it in a heartbeat.
The lack causes serious problems for a lot of tracking databases, and the suggestion to just use the next largest storage type isn't feasible for huge datasets.
US Constitution:
Article II: The Presidency
Section 1. The executive power shall be vested in a President of the United States of America. He shall hold his office during the term of four years, and, together with the Vice President, chosen for the same term, be elected, as follows:
This section is specifially separate from Article I (The Legislative Branch), and Article III (The Judiciary Branch).
(Yes, the vice president is mentioned in Article I, but specifically noted as having no vote)
The Vice President of the United States shall be President of the Senate, but shall have no vote, unless they be equally divided.
I realize this is nitpickly, but OC4 does not exist. They had several OC3s go down, and at one point he typoed OC4 in his blog. I find it amusing that one typo is what shows up on the Slashdot frontpage.
We've been using PoE for IP Telephony rollouts at corporate offices for years, for the exact reasons stated.
I suppose this is driving other uses of it as well, but PoE is neither new, nor unknown. Its just not an OMG kind of technology so it doesn't get a lot of airtime.
Particularly in installations where an existing PBX is being replaced with VoIP technology, its easy and cost effective to attach the call servers and PoE switches to existing UPS setups. Most companies have their IDF on UPS anyways, so it is fairly seamless and provides the same emergency features people have come to expect from PSTN and PBXes.
While there are a lot that are "better" (for certain sets of features), and cheaper, Cisco has the benefit of being the incumbent. While I realize they have a pseudo-monopoly on high end internetworking gear, I cannot afford to run the risk of incompatibility for the sake of buying something else. Frame is not always frame, to use a specific example, where interoperability across the link sometimes doesn't work as well as it should.
In addition to that, Cisco TAC is the best support service I have ever worked with. I'm sure I'll get a lot of horror stories about TAC in response, but at least as a partner I've had very very few negative experiences with TAC.
The simple fact is that in order to make your car faster, you need more air and more fuel. That's it.
Learn how your particular engine works, its tolerances, and its choke points. Fuel is easy to add more of, especially in a fuel injected car. Bigger injectors and the ability to change your fuel curve and monitor the engine status will fix that.
Air is harder. The MAF is often a choke point, as is the intake box. Throttle bodies will sometimes be an issue, although less so with the larger V-8 engines. For NA cars, you'll eventually need a stroker or something of that nature to get much more air (or adding forced induction).
When push comes to shove, an engine is a simple thing. Its just a big air pump. Learn the basics of engine management, fuel injection and timing, and static/dynamic compression if you're looking at forced induction, and the rest of it is easy.
Re:Where is the security measure? (was: Re:How lon
on
RFID MasterCard
·
· Score: 1
Assuming this technology is similar to the popular HID cards used for security, then its a combination of both. The reader is just a low-energy field at a particular frequency. The card has an indutor coil wound around the inside, connected to a small RF emitter with a 48-bit ID number (at least for HID, I would assume that number will be larger eventually).
When the coil passes through the specific frequency field that the reader is generating, it causes enough charge in the coil to send the ID in a burst to the reader. The larger the field, the further away they will work, however. This technology is being used in cars for auto-toll paying and such, which has a range of several feet, and some larger readers are stationed on either side of a doorway to read the cards of everybody passing through.
I whipped up a quick ebuild for Gentoo if anybody feels like running this on their computer. Just grab the files and drop them in a portage overlay via the usual methods. (You can find good documentation on using the overlay on gentoo's site)
As has been pointed out time and time again on NANOG and other operational mailing lists, DNS hijacking is still DNS hijacking, regardless of how noble the intent is.
From an operations standpoint, the impacts of Sitefinder are unfortunatly minimal now. Most of the major operational issues brought up when it was first released have been solved by either Verisign or by various application developers (ISC and other DNS developers) and are no longer an issue.
While I and many other people involved in operations agree that Sitefinder is a horrible idea ethically, nobody is helping their case with histronics and ad hominem attacks on Verisign's business practices, regardless of how true they are. All that does is gives Verisign more fuel for their "technocratic elite" arguments in press releases.
If you really want to fight this, tone down some of the passion and write to ICANN with legitimate concerns about the service and its effects. Crying foul about slimy business practices with no supporting evidence and a lot of sound and fury is a good way to make people who might be swayed agree with Verisign's claims of being attacked unjustly.
DB2 is, in my opinion, still a maturing product for anything that's not a mainframe or minicomputer. DB2 is a very robust RDBMS on S/390 and AS/400 systems, and has a lot of maturity there.
It's been available for AIX for quite some time, but it looks as if they didn't really put a lot of effor into it until recently, likewise the Linux version was quite buggy and had memory leaks until recently.
However, this appears to be changing rapidly now. Having just started using DB2 V 7.1, I can say I'm quite impressed. It's a fully functional, high-performance, and fully SQL compliant database. It may not perform as well as a well tuned Oracle database, but it is pretty close, and is far easier to get high performance out of.
In addition to that, the price is 1/4 or less than that of an equivalent Oracle license. So, I'm of the opinion that it will be growing quickly now that it's worth building on.
Now, as for comparing things like DB2 and Oracle (Which I think are probably the 2 highest-end RDBMSs out there) to Postgres and MySQL, there simply IS no comparison. Let's be honest, these companys haven't been getting away with charging what they do simply because of marketing. It's not like a small consumer product where people just buy it because it's there, these DBMs are researched for a long period of time be most companies before they are purchased.
I would never run what I classify as an Enterprise level database on any open source RDBMS. Not because I hate open source, but because it simply does not have the qualities that should be considered in an enterprised DB.
(An IBM developer already posted what those are, so I'm not going to repeat them.)
Slashdot Mirror
http://www.catalystframework.org/
Catalyst is mature, activly supported, and being used in a lot of locations. (Sixapart's Vox being the largest example)
So, no, we haven't been pushed out of the web market by other lauguages. Thanks for asking. =)
(There's Jifty too, but I have no person experience with it)
No, it will not. Peering is based on BGP, which doesn't do any sort of dynamic load balancing. Its effectivly a link-state protocol, as long as the peer with the best path (and other metrics that're not related to the current performance of the link) is still up, the traffic will prefer that peer, regardless of throughput.
However, peering agreements typically have clauses that prevent any sort of traffic shaping.
"A hundred thousand harmless junkies became criminals overnight, by Act of Congress, in nineteen twenty-seven. Ten years later, in thirty-seven, all the pot-heads in the country became criminals overnight, by Act of Congress. And they really were criminals, when the papers were signed. The guns prove it. Walk away from those guns, waving a joint, and refuse to halt when they tell you. Their Imagination will become your Reality in a second." -Illuminatus, Robert Anton Wilson
Something else.
Specifically, encoded network traffic data collected in a packed form of unsigned 16, 32 and 64 bit integers. While its possible to shift the numbers to be signed, it makes the supporting code substantially more complicated, since ultimatly most of these numbers are not counters, they represent identifiers, which means all queries and output would need to do these calculations as well. Certainly not impossible, but its a huge increase in code complexity.
If postgres supported native unsigned integer types, I'd switch to it in a heartbeat.
The lack causes serious problems for a lot of tracking databases, and the suggestion to just use the next largest storage type isn't feasible for huge datasets.
US Constitution:
Article II: The Presidency
Section 1. The executive power shall be vested in a President of the United States of America. He shall hold his office during the term of four years, and, together with the Vice President, chosen for the same term, be elected, as follows:
This section is specifially separate from Article I (The Legislative Branch), and Article III (The Judiciary Branch).
(Yes, the vice president is mentioned in Article I, but specifically noted as having no vote)
The Vice President of the United States shall be President of the Senate, but shall have no vote, unless they be equally divided.
I do, I do! I don't even wear ties, but I'd buy it just in case I needed to go to a funeral or something else formal.
I bet it would be a hit at weddings too.
Its the sea of Valusia, of course...
People who pirate software? Yeah, as a rule that's true.
People who USE pirated software? Eh, not so much. Its pretty obvious if you look at the comments sections on torrent sites.
I realize this is nitpickly, but OC4 does not exist. They had several OC3s go down, and at one point he typoed OC4 in his blog. I find it amusing that one typo is what shows up on the Slashdot frontpage.
We've been using PoE for IP Telephony rollouts at corporate offices for years, for the exact reasons stated.
I suppose this is driving other uses of it as well, but PoE is neither new, nor unknown. Its just not an OMG kind of technology so it doesn't get a lot of airtime.
Particularly in installations where an existing PBX is being replaced with VoIP technology, its easy and cost effective to attach the call servers and PoE switches to existing UPS setups. Most companies have their IDF on UPS anyways, so it is fairly seamless and provides the same emergency features people have come to expect from PSTN and PBXes.
Disclosure: I am a Cisco Channel Partner
While there are a lot that are "better" (for certain sets of features), and cheaper, Cisco has the benefit of being the incumbent. While I realize they have a pseudo-monopoly on high end internetworking gear, I cannot afford to run the risk of incompatibility for the sake of buying something else. Frame is not always frame, to use a specific example, where interoperability across the link sometimes doesn't work as well as it should.
In addition to that, Cisco TAC is the best support service I have ever worked with. I'm sure I'll get a lot of horror stories about TAC in response, but at least as a partner I've had very very few negative experiences with TAC.
The simple fact is that in order to make your car faster, you need more air and more fuel. That's it.
Learn how your particular engine works, its tolerances, and its choke points. Fuel is easy to add more of, especially in a fuel injected car. Bigger injectors and the ability to change your fuel curve and monitor the engine status will fix that.
Air is harder. The MAF is often a choke point, as is the intake box. Throttle bodies will sometimes be an issue, although less so with the larger V-8 engines. For NA cars, you'll eventually need a stroker or something of that nature to get much more air (or adding forced induction).
When push comes to shove, an engine is a simple thing. Its just a big air pump. Learn the basics of engine management, fuel injection and timing, and static/dynamic compression if you're looking at forced induction, and the rest of it is easy.
Assuming this technology is similar to the popular HID cards used for security, then its a combination of both. The reader is just a low-energy field at a particular frequency. The card has an indutor coil wound around the inside, connected to a small RF emitter with a 48-bit ID number (at least for HID, I would assume that number will be larger eventually).
When the coil passes through the specific frequency field that the reader is generating, it causes enough charge in the coil to send the ID in a burst to the reader. The larger the field, the further away they will work, however. This technology is being used in cars for auto-toll paying and such, which has a range of several feet, and some larger readers are stationed on either side of a doorway to read the cards of everybody passing through.
I whipped up a quick ebuild for Gentoo if anybody feels like running this on their computer. Just grab the files and drop them in a portage overlay via the usual methods. (You can find good documentation on using the overlay on gentoo's site)
The ebuild submission can be found here: http://bugs.gentoo.org/show_bug.cgi?id=49220
-Llarian
At least my mirror is running a very low capacity at the moment. I think people are figuring out that this is still in the "experimental" phase.
However, if you feel like using it, http://gentoo.llarian.net/experimental is one of the US Gentoo mirrors.
As has been pointed out time and time again on NANOG and other operational mailing lists, DNS hijacking is still DNS hijacking, regardless of how noble the intent is.
From an operations standpoint, the impacts of Sitefinder are unfortunatly minimal now. Most of the major operational issues brought up when it was first released have been solved by either Verisign or by various application developers (ISC and other DNS developers) and are no longer an issue.
While I and many other people involved in operations agree that Sitefinder is a horrible idea ethically, nobody is helping their case with histronics and ad hominem attacks on Verisign's business practices, regardless of how true they are. All that does is gives Verisign more fuel for their "technocratic elite" arguments in press releases.
If you really want to fight this, tone down some of the passion and write to ICANN with legitimate concerns about the service and its effects. Crying foul about slimy business practices with no supporting evidence and a lot of sound and fury is a good way to make people who might be swayed agree with Verisign's claims of being attacked unjustly.
There's a reason that DC->AC is called "inverting", whereas AC->DC is called "rectifying".
-Llarian
It's been available for AIX for quite some time, but it looks as if they didn't really put a lot of effor into it until recently, likewise the Linux version was quite buggy and had memory leaks until recently.
However, this appears to be changing rapidly now. Having just started using DB2 V 7.1, I can say I'm quite impressed. It's a fully functional, high-performance, and fully SQL compliant database. It may not perform as well as a well tuned Oracle database, but it is pretty close, and is far easier to get high performance out of.
In addition to that, the price is 1/4 or less than that of an equivalent Oracle license. So, I'm of the opinion that it will be growing quickly now that it's worth building on.
Now, as for comparing things like DB2 and Oracle (Which I think are probably the 2 highest-end RDBMSs out there) to Postgres and MySQL, there simply IS no comparison. Let's be honest, these companys haven't been getting away with charging what they do simply because of marketing. It's not like a small consumer product where people just buy it because it's there, these DBMs are researched for a long period of time be most companies before they are purchased.
I would never run what I classify as an Enterprise level database on any open source RDBMS. Not because I hate open source, but because it simply does not have the qualities that should be considered in an enterprised DB. (An IBM developer already posted what those are, so I'm not going to repeat them.)
-Llarian