That's one of the stupidest things I've heard in decades. Do I have to be a union member to steal stuff? *shakes head*
Nothing is ever going to stop piracy ("theivery".) People have bootleged crap for as long as it's been possible. People used to bootleg vinyl LP's. People still bootleg audio and video tapes. And these things take real time to duplicate. If anything, modern technologies have made the process cheaper, faster, and easier. Back in the 80's, piracy was a limited thing as the tech was slow (300-2400 baud modems), clunky (you have to run in the right circles), expensive, and hard-to-find. Today, it's the complete opposite... high-speed internet access is "everywhere" (except my parents house *grin*, TW still won't extend cable to it); 40x(cd)/16x(DVD) burners are in the 50$ range and available everywhere computers are sold; software for "one-click" duplication of just about anything is everywhere -- and probablly supplied with the drive if not sold on the same shelf.
No, it is not. It is, however, a very difficult and time consuming task to recover. You don't see r/w errors because a) the heads on your drive aren't that sensitive (we're talking very, VERY, small differences that are often not *directly* readable -- differiential magnetic analysis), and b) modern drives use highly complex error correction methods that are all but invisable under normal use.
It's *alot* more involved than simple pencil-and-paper inspection where you can instantly see where a "1" has been erased and a "0" written in it's place. It's sorta like finding out what's under a blob of White Out(tm) when the White Out(tm) -- and what's under it -- isn't even noticable.
["fast fills"? I think you have "hard drive" and "eeprom" confused. And even then, you've got it backwards... eeprom's read 0xff when empty; 1's are free, 0's have to be burned.]
I would like to see higher upload speeds because it's really annoying to try and telecommute at 384k...
I've done this (over IPSec that eats even more BW) and 384k is perfectly acceptable. I've also done so via 64k and 128k (ISDN.) 64k is painfully slow; I've seen better and won't go back:-) 128k is ok, but it takes a smart terminal app (citrix, rdp, vnc, etc.) to avoid unnecessary repainting of large areas -- citrix is the only thing I've seen work well in low bandwidth and/or congestion.
(Note: don't connect to your windows desktop ("terminal server", "remote desktop", etc.) at 24bit... 8bit color depth is all you need.)
It's easy for the cable network to increase the download speeds as that's what the network is designed for. Each channel supports 30Mbps down/10Mbps up. Flooding the downstream channel(s) isn't a major problem as everything will level out -- those using the most bits will see the biggest hit. However, flooding the upstream channel(s) will kill everyone... I don't get anymore packets until I ACK the one's I've received. And QoS policies don't work upstream -- you cannot shape the bits until you have them, and in the upstream direction, once you have the bits, it's already too late; the damage is done.
TW hands out SB5100's now (have been for some time now.) However, there are still an ass load of older SB's out there (even 2100's, 'tho rare.) I know of a several 3100's and 4100's around here.
Personally, I like the 5100... nice tiny black thing with blinkin' LED(s). It's supposed to be the "hacker proof" version, but we all know there's no such thing. (people hack cable boxes that are filled with black tar...)
You have it backwards... the torrent file is the pointer to illegal goodies. The tracker has zero idea what you're trading; it only has the info_hash to tell things apart. So, to what infringing material does "B093EBF1 EAA7BF6D 1DC12BC7 6580E724 B2458208" pertain? That and maybe the filesize is all the tracker has to go on.
That's the difference between "suspecting" a thing and "actual knowledge of" a thing. I may suspect my neighbor is selling illegal drugs, but unless I go buy some from him, I don't have actual knowledge that he's selling illegal drugs -- of course, that would also mean I'd've broken a few laws too:-) Here, it's not enough for them to suspect many of the torrents on the site are for trading copyrighted material from unauthorized distributors. They would have to download at least part of each torrent to verify that it's a) copyrighted content (which, btw, everything is), and b) being distributed by those not authorized to do so. I doubt any torrent site on the 'net verifies the content to which the torrents pertain. (the torrent itself is not a copyright violation.)
Well, yes and no. US law has the notion of being an accessory to a crime or "aiding and abetting". So, telling me were to buy bootleg CDs and DVDs with a reasonable expectation that I'll act on that information makes you an accessory to my crimes. I don't know of anyone ever being arrested for something so lame, but it's within the letter of the law.
The MPAA and RIAA can (successfully) argue that providing torrents to their copyrighted content is inducing a crime. These torrent sites are facilitating the commission of a crime. ("aiding and abetting" in the legal tongue) The tracker itself is a darker area as the tracker has no knowledge of what it's tracking; all it needs is the info_hash. Most trackers know more than that, but that's all it needs.
I cannot speak about what might be legal or illegal all over the world, but in the US, any cases brought against torrent or tracker sites have a very good chance of loosing. Their greatest power is in frightening people (see also: suprnova.org), because very few have the money and time to go to court (for what might take years.)
[Hmm, can one even be an accessory to a civil charge?]
Knowing where to buy illegal drugs is not against the law. And technically, telling someone this information is also not against the law. However, it can be argued that doing so is "aiding and abetting" which makes you an accessory to a crime. It's the same argument that gets gun dealers "questioned"... one cannot (legally) sell someone a gun knowing they're going to (imminently) commit a crime with it. It's a legal grey area as just about everything ever made can, in some fashion, be used to break numerous laws. (What's the point of making cars that can go 120+ mph when the max. speed anywhere in the country is 75? The car maker is "aiding and abetting" my speeding crimes.)
The tracker itself, having no copyright'd content, is not breaking any laws in most parts of the world.
how the f*** do you "infect the torrents"? These things aren't word documents with macros; they are strictly formated files with no executable data in them. The client WILL ignore fields it doesn't recognize. There's no chance of buffer overflows (unless the client is really shitty) as every field has the size specified in the torrent.
Please take a few minutes to understand the technology before repeating the BS rumors of the insane and criminally stupid.
[Btw, that torrent is usable anywhere... linux, windows nt 4.0 on an alpha, BeOS on BeBox, VxWorks (the RTOS inside your cable modem),...]
There are 2 obvious problems with port 25 redirects... 1) all it takes is one idiot to get the ISP's mail server(s) blacklisted, 2) any problem with the ISP's mail server(s) gives you no other option for sending email.
I've seen #1 so many times I've given up on the whole process. You cannot black list an ISP's mail server(s) because a handful of junk went through it -- that junk is less than 1% of what's moving through there. How many times have AOL's servers been blacklisted? I stopped counting years ago.
#2 is a serious pain in the ass. No ISP's mail farm can deal with my email faster or more reliablly than my own computers from my own connection. RR uses (used?) Microsloth's SMTP crap. I don't know what Earthlink uses, but it's much slower than handling it myself. Network Solutions (selling mail service for businesses!?) uses Mirapoint systems so mis-managed and overloaded, it can take most of a day for a single message to get through it (if at all.)
Most dynamic address ranges are already listed in various anti-spam blocklists. That works far, FAR better than port 25 redirection -- esp. when no one informs you of the trap (Cox, Rogers, The Hilton,...)
Not all of them were T1's. I know for a fact he had at least 2 DS3's. (because I saw the router they went into.) The ironic thing is, spammers always pay the bills on time.
This is incorrect. As one who has dealt with Lucent 5ESS switches, it's as "easy" to turn off as it is to turn on in the first place. It's one of the many line provisioning options.
Now, I say "easy" as the term is certainly relative when working with telco switches. I won't bore people with stories; suffice to say the CLI is very cryptic and the menu interface (from which all real work is done) is a bit complicated to the uninitiated.
The idea is that in order to write data to any sector on one of the drives, the sectors from six of the other drives need to be read, all XOR'd together, and then the result written to the remaining drive
That's not RAID 5. That's RAID 3 or 4. 5 is striped data AND parity. What you are describing has a single parity drive. (And yes, it's pretty damned slow.) With the additional caching effects, if you breifly delay writing, both blocks of a parity set may get written at the same time, so zero block reads, 3 block writes. Normally, there will be 1 block read and 2 blocks written.
Earthlink isn't a cable company. They are an ISP, but they don't run the HFC network. So, who's cable network are you on? That determines what gets filtered. I have an Earthlink (mindspring) IP but it's all within RR's network. My packets mingle with every other ISP-reseller's packets.
Indeed. Fiber is full-duplex and synchronous. What's the point of making it async? And what's with the PPPoE spew? It's not like your neighbor is gonna steel Fios service from you without them knowing. (The same arguement goes for DSL.)
Haven't they heard of.1x?
(Bandwidth ain't free. And high bit rate optical gear is very expensive.)
Yeah, that stupid. Service availablity is based on physical location but the monkeys ask for a f'ing phone number. How 'bout asking for a freakin' address you idiots?
TCP Intercepts have a high-water mark where uncompleted connections are reaped. For Cisco IOS and PixOS, those thresholds are configurable -- and they work if you can stand the volume of packets.
You honestly expect to prosecute 10,000 hacked windows box owners? (I'd contribute to that legal fund:-))
interface FastEthernet0/0...
ip verify unicast source reachable-via...
Cisco started RPF (reverse path forwarding) a number of years ago. It uses the CEF (Cisco Express Forwarding) table's FIB (Forwarding Information Base) to know if the packet came from where it should've. Since this is within the packet's normal switching path, the check is practically free.
It doesn't work like that. BGP doesn't specify a destination for a prefix -- it's more a "give me X" type thing; the destination is the peer announcing the prefix. Likewise, you cannot easily punch a hole in the middle of your netblock with a null route. (in fact, the null route is how you ensure the netblock is always announced.) Add to that all the peering filters and the minimum prefix length (/20, btw), and "it's just not gonna happen". That traffic is going to show up at the ISP along with all the traffic you really do want. There's no way to get the entire internet to filter it for you. That's why the DDoS is such an annoying bugger.
My apartment leaks a few degF per hour, but it was built 8 years ago by a bunch of half-drunk mexicans in about three months. (there are beer cars inside some of the walls!) The windows are the only things that properly seal -- because they're pre-manufactured, triple pane windows. I plug the doors with plastic grocery bags but it bleeds heat rapidly.
As you already have a gas hookup, why not get a natural gas powered generator? They aren't that much more expensive.
If they don't wish to transit traffic across their network, they really aren't legally obligated to.
Wrong. The instant an ISP makes distinctions between the bits flyin' around, it loses common carrier status and all the legal protections of being a common carrier (i.e. "we ain't responsible for the shit our customers spew") And in that instant, RIAA and MPAA monkeys fly in to sue the hell out of them for all the copyright infringements of their clients, for which they are now responsible.
ISP's that install transparent proxies for things like http and smtp are on very thin ice. But tend to get away with it without challenge.
Comcast, Adelphia and others did not use the public money to build [their network]
Yes and no. While there certainly wasn't a public bond issue to build their networks, taxpayer money was involved. Many cable operators are paid by city goverments to build and service the city's cable system. The cable industry (in the US) is a licenced monopoly. I know of no areas of the US where there are multiple cable operators. Cable lines are "public services" even if they haven't been tarrifed as such, yet.
It is worth noting, the PSTN was built with taxpayer money as well -- in almost exactly the same way. The only difference is that the PSTN was built many decades ago. Until the various PUC's declare cable network lines as public goods, they will have their relatively closed monopoly.
ISPs around the States sued
And while they didn't "win" in the sense of getting the cable networks declared public infrastructure, they did poke holes in the monopoly... the cable companies are required to openly resell access to the network. They still own the wiring and hardware, but they don't have a monopoly on ISP service. People like AT&T, Earthlink, and AOL all sell broadband within various cable provider domains.
Comcast can do whatever they please
They can get away with a lot of things Public Utilities can't. But that's not to say they can do anything they want. They are still a monopoly. And they are still bound by entire libraries of commerce laws. That's what this ultimately boils down to... as an ISP, they cannot filter or restrict traffic (i.e. make VoIP only work for them) without losing common carrier status and all the legal protections that afords.
(No one has evoked the Sherman Act. That's much more powerful than restraint of trade.)
It's no more a server than an IM client. Servers accept connections from anything without further assistance -- web servers, ftp servers, irc servers, dns servers... they don't rely on a gateway to mediate the connection between both endpoints; the server is one of the endpoints. Your DNS server doesn't notify your web server that xxx has asked it for {service = http; host = www.foo.com.} (for starters DNS doesn't work that way. and I know of no applications that will ask for (very rare) service records from DNS.)
An ATA is not a server. It's registered with/connected to a specific call manager. The call manager is what brokers incoming and outgoing calls. The ATA is not directly addressible by anything other than the call manager and the end-point(s) eventually directed to it by the call manager.
Just because I know the IP address of your Cisco IP phone doesn't mean I can make it ring (without spoofing the call manager.)
Slashdot Mirror
- Piracy should not be for the masses
That's one of the stupidest things I've heard in decades. Do I have to be a union member to steal stuff? *shakes head*Nothing is ever going to stop piracy ("theivery".) People have bootleged crap for as long as it's been possible. People used to bootleg vinyl LP's. People still bootleg audio and video tapes. And these things take real time to duplicate. If anything, modern technologies have made the process cheaper, faster, and easier. Back in the 80's, piracy was a limited thing as the tech was slow (300-2400 baud modems), clunky (you have to run in the right circles), expensive, and hard-to-find. Today, it's the complete opposite... high-speed internet access is "everywhere" (except my parents house *grin*, TW still won't extend cable to it); 40x(cd)/16x(DVD) burners are in the 50$ range and available everywhere computers are sold; software for "one-click" duplication of just about anything is everywhere -- and probablly supplied with the drive if not sold on the same shelf.
No, it is not. It is, however, a very difficult and time consuming task to recover. You don't see r/w errors because a) the heads on your drive aren't that sensitive (we're talking very, VERY, small differences that are often not *directly* readable -- differiential magnetic analysis), and b) modern drives use highly complex error correction methods that are all but invisable under normal use.
It's *alot* more involved than simple pencil-and-paper inspection where you can instantly see where a "1" has been erased and a "0" written in it's place. It's sorta like finding out what's under a blob of White Out(tm) when the White Out(tm) -- and what's under it -- isn't even noticable.
["fast fills"? I think you have "hard drive" and "eeprom" confused. And even then, you've got it backwards... eeprom's read 0xff when empty; 1's are free, 0's have to be burned.]
Time to add a copyright notice to that MD5 signed config file :-)
Of course, the DMCA has little to do with copyright in actual practice. Saddly.
- I would like to see higher upload speeds because it's really annoying to try and telecommute at 384k...
I've done this (over IPSec that eats even more BW) and 384k is perfectly acceptable. I've also done so via 64k and 128k (ISDN.) 64k is painfully slow; I've seen better and won't go back(Note: don't connect to your windows desktop ("terminal server", "remote desktop", etc.) at 24bit... 8bit color depth is all you need.)
It's easy for the cable network to increase the download speeds as that's what the network is designed for. Each channel supports 30Mbps down/10Mbps up. Flooding the downstream channel(s) isn't a major problem as everything will level out -- those using the most bits will see the biggest hit. However, flooding the upstream channel(s) will kill everyone... I don't get anymore packets until I ACK the one's I've received. And QoS policies don't work upstream -- you cannot shape the bits until you have them, and in the upstream direction, once you have the bits, it's already too late; the damage is done.
TW hands out SB5100's now (have been for some time now.) However, there are still an ass load of older SB's out there (even 2100's, 'tho rare.) I know of a several 3100's and 4100's around here.
Personally, I like the 5100... nice tiny black thing with blinkin' LED(s). It's supposed to be the "hacker proof" version, but we all know there's no such thing. (people hack cable boxes that are filled with black tar...)
You have it backwards... the torrent file is the pointer to illegal goodies. The tracker has zero idea what you're trading; it only has the info_hash to tell things apart. So, to what infringing material does "B093EBF1 EAA7BF6D 1DC12BC7 6580E724 B2458208" pertain? That and maybe the filesize is all the tracker has to go on.
That's the difference between "suspecting" a thing and "actual knowledge of" a thing. I may suspect my neighbor is selling illegal drugs, but unless I go buy some from him, I don't have actual knowledge that he's selling illegal drugs -- of course, that would also mean I'd've broken a few laws too :-) Here, it's not enough for them to suspect many of the torrents on the site are for trading copyrighted material from unauthorized distributors. They would have to download at least part of each torrent to verify that it's a) copyrighted content (which, btw, everything is), and b) being distributed by those not authorized to do so. I doubt any torrent site on the 'net verifies the content to which the torrents pertain. (the torrent itself is not a copyright violation.)
(The "letter of the law" can be such a b***h.)
Well, yes and no. US law has the notion of being an accessory to a crime or "aiding and abetting". So, telling me were to buy bootleg CDs and DVDs with a reasonable expectation that I'll act on that information makes you an accessory to my crimes. I don't know of anyone ever being arrested for something so lame, but it's within the letter of the law.
The MPAA and RIAA can (successfully) argue that providing torrents to their copyrighted content is inducing a crime. These torrent sites are facilitating the commission of a crime. ("aiding and abetting" in the legal tongue) The tracker itself is a darker area as the tracker has no knowledge of what it's tracking; all it needs is the info_hash. Most trackers know more than that, but that's all it needs.
I cannot speak about what might be legal or illegal all over the world, but in the US, any cases brought against torrent or tracker sites have a very good chance of loosing. Their greatest power is in frightening people (see also: suprnova.org), because very few have the money and time to go to court (for what might take years.)
[Hmm, can one even be an accessory to a civil charge?]
Knowing where to buy illegal drugs is not against the law. And technically, telling someone this information is also not against the law. However, it can be argued that doing so is "aiding and abetting" which makes you an accessory to a crime. It's the same argument that gets gun dealers "questioned"... one cannot (legally) sell someone a gun knowing they're going to (imminently) commit a crime with it. It's a legal grey area as just about everything ever made can, in some fashion, be used to break numerous laws. (What's the point of making cars that can go 120+ mph when the max. speed anywhere in the country is 75? The car maker is "aiding and abetting" my speeding crimes.)
The tracker itself, having no copyright'd content, is not breaking any laws in most parts of the world.
how the f*** do you "infect the torrents"? These things aren't word documents with macros; they are strictly formated files with no executable data in them. The client WILL ignore fields it doesn't recognize. There's no chance of buffer overflows (unless the client is really shitty) as every field has the size specified in the torrent.
Please take a few minutes to understand the technology before repeating the BS rumors of the insane and criminally stupid.
[Btw, that torrent is usable anywhere... linux, windows nt 4.0 on an alpha, BeOS on BeBox, VxWorks (the RTOS inside your cable modem),...]
There are 2 obvious problems with port 25 redirects... 1) all it takes is one idiot to get the ISP's mail server(s) blacklisted, 2) any problem with the ISP's mail server(s) gives you no other option for sending email.
...)
I've seen #1 so many times I've given up on the whole process. You cannot black list an ISP's mail server(s) because a handful of junk went through it -- that junk is less than 1% of what's moving through there. How many times have AOL's servers been blacklisted? I stopped counting years ago.
#2 is a serious pain in the ass. No ISP's mail farm can deal with my email faster or more reliablly than my own computers from my own connection. RR uses (used?) Microsloth's SMTP crap. I don't know what Earthlink uses, but it's much slower than handling it myself. Network Solutions (selling mail service for businesses!?) uses Mirapoint systems so mis-managed and overloaded, it can take most of a day for a single message to get through it (if at all.)
Most dynamic address ranges are already listed in various anti-spam blocklists. That works far, FAR better than port 25 redirection -- esp. when no one informs you of the trap (Cox, Rogers, The Hilton,
Not all of them were T1's. I know for a fact he had at least 2 DS3's. (because I saw the router they went into.) The ironic thing is, spammers always pay the bills on time.
This is incorrect. As one who has dealt with Lucent 5ESS switches, it's as "easy" to turn off as it is to turn on in the first place. It's one of the many line provisioning options.
Now, I say "easy" as the term is certainly relative when working with telco switches. I won't bore people with stories; suffice to say the CLI is very cryptic and the menu interface (from which all real work is done) is a bit complicated to the uninitiated.
- The idea is that in order to write data to any sector on one of the drives, the sectors from six of the other drives need to be read, all XOR'd together, and then the result written to the remaining drive
That's not RAID 5. That's RAID 3 or 4. 5 is striped data AND parity. What you are describing has a single parity drive. (And yes, it's pretty damned slow.) With the additional caching effects, if you breifly delay writing, both blocks of a parity set may get written at the same time, so zero block reads, 3 block writes. Normally, there will be 1 block read and 2 blocks written.Earthlink isn't a cable company. They are an ISP, but they don't run the HFC network. So, who's cable network are you on? That determines what gets filtered. I have an Earthlink (mindspring) IP but it's all within RR's network. My packets mingle with every other ISP-reseller's packets.
Surprisingly, nothing is being filtered.
Indeed. Fiber is full-duplex and synchronous. What's the point of making it async? And what's with the PPPoE spew? It's not like your neighbor is gonna steel Fios service from you without them knowing. (The same arguement goes for DSL.)
.1x?
Haven't they heard of
(Bandwidth ain't free. And high bit rate optical gear is very expensive.)
Yeah, that stupid. Service availablity is based on physical location but the monkeys ask for a f'ing phone number. How 'bout asking for a freakin' address you idiots?
Nah. Take over the channel. set "+v" and sit back and enjoy the silence. He won't be able to command the bots if he cannot speak to them.
Of course, he could do the same. And he'd still see all the bots entering and leaving the channel...
TCP Intercepts have a high-water mark where uncompleted connections are reaped. For Cisco IOS and PixOS, those thresholds are configurable -- and they work if you can stand the volume of packets.
:-))
You honestly expect to prosecute 10,000 hacked windows box owners? (I'd contribute to that legal fund
It's anctually even easier than that...
... ...
interface FastEthernet0/0
ip verify unicast source reachable-via
Cisco started RPF (reverse path forwarding) a number of years ago. It uses the CEF (Cisco Express Forwarding) table's FIB (Forwarding Information Base) to know if the packet came from where it should've. Since this is within the packet's normal switching path, the check is practically free.
[See Also: Configuring Unicast Reverse Path Forwarding]
(As others have stated, this will not stop a DDoS as they aren't spoofed.)
It doesn't work like that. BGP doesn't specify a destination for a prefix -- it's more a "give me X" type thing; the destination is the peer announcing the prefix. Likewise, you cannot easily punch a hole in the middle of your netblock with a null route. (in fact, the null route is how you ensure the netblock is always announced.) Add to that all the peering filters and the minimum prefix length (/20, btw), and "it's just not gonna happen". That traffic is going to show up at the ISP along with all the traffic you really do want. There's no way to get the entire internet to filter it for you. That's why the DDoS is such an annoying bugger.
My apartment leaks a few degF per hour, but it was built 8 years ago by a bunch of half-drunk mexicans in about three months. (there are beer cars inside some of the walls!) The windows are the only things that properly seal -- because they're pre-manufactured, triple pane windows. I plug the doors with plastic grocery bags but it bleeds heat rapidly.
As you already have a gas hookup, why not get a natural gas powered generator? They aren't that much more expensive.
- If they don't wish to transit traffic across their network, they really aren't legally obligated to.
Wrong. The instant an ISP makes distinctions between the bits flyin' around, it loses common carrier status and all the legal protections of being a common carrier (i.e. "we ain't responsible for the shit our customers spew") And in that instant, RIAA and MPAA monkeys fly in to sue the hell out of them for all the copyright infringements of their clients, for which they are now responsible.ISP's that install transparent proxies for things like http and smtp are on very thin ice. But tend to get away with it without challenge.
- Comcast, Adelphia and others did not use the public money to build [their network]
Yes and no. While there certainly wasn't a public bond issue to build their networks, taxpayer money was involved. Many cable operators are paid by city goverments to build and service the city's cable system. The cable industry (in the US) is a licenced monopoly. I know of no areas of the US where there are multiple cable operators. Cable lines are "public services" even if they haven't been tarrifed as such, yet.It is worth noting, the PSTN was built with taxpayer money as well -- in almost exactly the same way. The only difference is that the PSTN was built many decades ago. Until the various PUC's declare cable network lines as public goods, they will have their relatively closed monopoly.
- ISPs around the States sued
And while they didn't "win" in the sense of getting the cable networks declared public infrastructure, they did poke holes in the monopoly... the cable companies are required to openly resell access to the network. They still own the wiring and hardware, but they don't have a monopoly on ISP service. People like AT&T, Earthlink, and AOL all sell broadband within various cable provider domains.- Comcast can do whatever they please
They can get away with a lot of things Public Utilities can't. But that's not to say they can do anything they want. They are still a monopoly. And they are still bound by entire libraries of commerce laws. That's what this ultimately boils down to... as an ISP, they cannot filter or restrict traffic (i.e. make VoIP only work for them) without losing common carrier status and all the legal protections that afords.(No one has evoked the Sherman Act. That's much more powerful than restraint of trade.)
It's no more a server than an IM client. Servers accept connections from anything without further assistance -- web servers, ftp servers, irc servers, dns servers... they don't rely on a gateway to mediate the connection between both endpoints; the server is one of the endpoints. Your DNS server doesn't notify your web server that xxx has asked it for {service = http; host = www.foo.com.} (for starters DNS doesn't work that way. and I know of no applications that will ask for (very rare) service records from DNS.)
An ATA is not a server. It's registered with/connected to a specific call manager. The call manager is what brokers incoming and outgoing calls. The ATA is not directly addressible by anything other than the call manager and the end-point(s) eventually directed to it by the call manager.
Just because I know the IP address of your Cisco IP phone doesn't mean I can make it ring (without spoofing the call manager.)