Posted by
CmdrTaco
on from the but-when-do-we-get-it-on-I2 dept.
Ant noted that there are
now quake games running on IPv6. Now once we get it running on I2, and someone manages to bring these 2 critical technologies to my bedroom, the world will be
a better place.
I'm not impressed. Give my IPv6 FIRST, then get me the games that can support it. Don't be putting the cart before the horse people. Until the ISP's are supporting IPv6, no one is going to care about this stuff. It's a fun project sure, but it's not even being used on an up-to-date game!
Re:Quake on IPv6. So what?
by
Fishstick
·
· Score: 2
Damn, the cup's half empty!
>but it's not even being used on an up-to-date game!
This is probably because the full source is available for Quake1.
--
There is much cruelty in the universe, John. Yeah, we seem to have the tour map.
I have no clue - could someone explain what happens when I do a lookup on that hostname and get 200.132.73.102 (on my horrid little IPv4 DNS server...)
And I just don't think hex sextuplets look as good - though they sound better:-)
Re:Quake on IPv6. So what?
by
Erasmus+Darwin
·
· Score: 3
Give my IPv6 FIRST, then get me the games that can support it.
But getting you IPv6 is more likely to happen when more software supports it. And software is more likely to support it as it becomes more likely that IPv6 will be accepted. In short, it's a chicken-and-egg problem. Anything people can do try drum up support, acceptance, and publicity for IPv6 is going to help. There's no magic switch that can be flipped to make everything IPv6 overnight.
Stop Talking, Start Doing
by
The-Pheon
·
· Score: 5
Everyone is saying "We must have ipv6 before we should worry about
games". The truth is that you can set yourself up with ipv6 right
now! So stop talking and start doing. you can get directions here.
I don't have to have ported a game to IPv6 to be unimpressed. That's the great part about opinions.
As for getting it myself, I'm talking about a non-tunneled version of it. Until an ISP hands me an IPv6 addy, I'm not interested. There's no REASON for me to "do it myself" yet.
Re:Quake on IPv6. So what?
by
Brandon+Hume
·
· Score: 3
Several things:
That version of Quake was put up by Viagenie a long time ago. I'm assuming the Slashdotters are calling it "new" just because it was added to the freenet6.net webpage with "New!" written next to it. ("New" refers to the LINK being new, not the game, or even the port!)
Secondly, in order to port Quake to IPv6, they needed Quake source. I'm sure if they'd had access to Quake 3 source at the time, they would have used that. But most game companies don't give away the source to their current money-makers, for some odd reason.
Third, while ISPs dragging their feet getting IPv6 to their customers IS getting annoying, you don't have to wait for them; that's what services like freenet6.net are for. There'll be lag problems due to the tunnelling, but you'll at least have something to play with.
And as for putting the cart before the horse - no one is going to switch to IPv6 if there aren't any programs that take advantage of it. If you try to get IPv6 everywhere before anything uses it, it won't happen. Yes, IPv6 is a superior protocol. That doesn't mean anything. Its human nature (not to mention SOP at ISPs) to bailing-wire and band-aid things to the point of destruction before going through the "work" of an upgrade.
Thankfully, I have native IPv6 access between my workstation, my machines at home, and Ca*Net3, so I'm ready to go.:) --
Brandon Hume
hume -> BOFH.Halifax.NS.Ca, http://WWW.BOFH.Halifax.NS.Ca/
I can delegate/64's to your home or office network if you are looking to get connected to the 6Bone on your Linux/BSD/Windows boxes.
I have dual redundant tunnels to cisco and Sprintlink via a fractional 10Mb ATM DS-3.
Feel free to drop me an e-mail if you are interested. I'd be more then happy to help
you get connected. (- the spam-me)
-Pat
NetBSD IPv6-enabled Quake package
by
jmcneill
·
· Score: 2
IPv6 should be supported natively in Whistler/Windows XP. Apparently, the word is that their stack will be dual v4/v6 and has the capability to translate (not 6to4) IPv4 socket calls to IPv6.
Their stack is pretty sweet, it supports IPv6 forwarding as a router, tunneling, 6to4.. regardless of whether you are running Professional or Server. They also have rudimentary MobileIP support and a beta web based tunnel broker that interfaces with Cisco routers.
No ISP'll implement it without demand. No demand without any software. No software without any ISP.
IPv6 Quake offers a (partial) solution to this. If it raises the whispers for IPv6 to at least a rumbling, it'll have done amazing good for the Internet as a whole.
-- It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
For reasonably fast networks such as a 100MB LAN, most of the latency comes from the software, i.e. the time it takes the operating system to get the data from an application and out onto the network card.
See Tanenbaum's "Computer Networks", 3rd edition, if you are interested in learning more, its a good introduction to networking in general. From pg 562: "Rule #1, CPU speed is more important than network speed". "Long experience has shown that in nearly all networks, operating system and protocol overhead (emphasis mine) dominates actual time on the wire." He goes on to demonstrate that although the theoretical minimum delay getting a frame out onto an 802.3 is 102 microsecs, in practice, it can take up to 15 times longer than that just to get the frame out - the delay is in the software, and protocols make up a large part of the problem.
It isn't about bandwidth. Networks are measured by two things: bandwidth and latency, or the product of these, "bandwidth latency product". For a network to perform well both of these things needs to be good, but on most current networks the bottleneck is not the bandwidth (especially for a game like Quake, where there are lots of small packets, so improving the latency improves the feel of the game more readily than bandwidth improvements.)
Its not uncommon for a packet to be copied in memory 4 or more times from when the application sends it until it gets out onto the wire. Quake uses UDP, which sits on IP, which gets wrapped into an 802.3 frame on a LAN (after doing some arp cache checking of course), and as well as being wrapped up several times, extra copying is also involved in passing the data from the app down to the drivers (and normally within an application itself). All this crap takes time (see "Rule #4: Minimize copying"). Optimising the protocols and software can have a very noticeable effect on network performance. Windows98SE's TCP/IP stack is noticably slower than NT's (I've done some tests copying large numbers of files, typically 30 to 50 % quicker) - this is why they used NT's TCP stack in WinMe, and its also why John Carmack was itching to work on the Linux IP stack a few months ago, to try prod MS into doing something about the lousy performance of Win98's networking. "Rule #5: YOu can buy more bandwidth but not lower delay".
The design and optimization of protocols has a big influence on network performance.
Re:Protocols and latency
by
sg_oneill
·
· Score: 2
it's interesting actually that this ties in with quake and many fps (half-life springs to mind here) that quite often frame rate makes a huge difference on network lag in the game. This is quite counter-intuitive, but the reason is, is that many games poll the network between frames to update state data. Thus at 50 fps, the network is polled 50 times a second. At 25 fps the server is half as responsive
-- Excuse the Unicode crap in my posts. That's an apostrophe, and slashdot is busted.
Re:Protocols and latency
by
BeanThere
·
· Score: 2
I've done a fair amount of network programming so I'm not completely clueless here.. anyway, I've found that sometimes (depending on how the network code is structured) a higher frame rate can make the network less responsive. Games like Quake are frame-rate limited (even at 90fps, on a good PC, you'll have some idle). Limitting the frame rate implies having a bit of time left over after each frame leaves some CPU idle which (depending on the OS too) can give a quite noticeable boost to the performance on a fast network such as a 100MB LAN. If the rendering process/thread is hogging as much CPU time as it can (3d rendering tends to be very CPU intensive), processing of network packets tends to get worse. This is usually most noticeable if the network communications makes use of a seperate thread or process, but I don't think Quake is structured like that - Quake seem to work as you say, polling the network between frames. The network code I've written for our work uses a seperate thread, and limiting the frame rate does help a little. Other network systems that I've done some system integration with, which use a seperate process for network communications on WinNT, leaving some CPU idle gave a VERY noticeable boost to network performance.
of course, very slow frame rates (like 25 fps) will again be counterproductive to the network, for games like Q3, as you say. I've found Q3 quite interesting to study, they seem to use 100% of the CPU no matter what the frame limit is - its 100% even if you limit the frame rate to 10. Presumably they continue to consume all CPU time with the game simulation and network.
A LOT of quake is already on I2
by
Tom7
·
· Score: 3
Lots of internet Quake is played at universities, and traffic between Internet2 schools already passes over I2 backbones instead of the "regular" internet. For instance, anybody with a hookup to the Abilene backbone will get great pings to my Q3F server. Unfortunately, your bedroom is probably not considered an academic institution, so you might have a tough time convincing the I2 people to connect you in, IPv6 or not!
Quake1 is GPL, if i remember correctly. Of course, they'll still be wanting to actually make some money off Quake3, but they'll probably open that up in a few years, when Carmack's newest creations are making Q3 look old.
Re:I expect it will make no difference at all.
by
WolfWithoutAClause
·
· Score: 2
Lag is caused by three things:
a) Einstein (well actually maxwell's laws) i.e. speed of light delay. But this is usually quite small e.g. UK to East Coast US is about 3000km i.e. maybe 30 milliseconds delay allowing for the refractive index of fiber
b) Hops (each machine adds a certain minimum delay, depending on the hardware/software on that host)
c) heavy traffic (when there are lots of packets, the lengths of queues increase, and introduce delays)
Anyway the point is that you can't do anything about a), [short of digging a straight line between the two sites(!), or moving yourselves closer together], b) you might change with extreme difficulty if at all.
However c) is a variable delay and relates to the lack of quality of service guarantees by the network. The kicker here is that the IP6 has protocols that can give guarantees (excepting network failures once the connection is set up).
c) is probably dominant or important in quite a lot of situations, and yes, IP6 does or atleast can help.
Also, in a Quake scenario, any reduction in variation translates into a big help. Variable delays can make it very difficult to play. Playing Quake with what amounts to a higher priority link compared to web page downloads or ftp sessions is a godsend - it isn't as if Quake uses much bandwidth, but it is very timing sensitive.
The average ISP runs almost all Cisco kit - even IOS 12.1(5)T, one of the most recent ones, does not support IPv6, and most ISPs will be running earlier versions. I believe IPv6 support from Cisco (in non-beta IOS versions) will be out this year. I believe Bay and many other routers already support IPv6, but that doesn't matter for most ISPs.
There's also a lot of planning required in how exactly the ISP should switch on IPv6 in their networks - their whole routing setup and DNS infrastructure needs extending, for a start, not to mention the various 6-to-4 conversion schemes. It's not just a matter of turning IPv6 on.
When IPv6 is enabled, you should not have to type 128 bit addresses in any case (did you type in the address for slashdot.org, or did you use DNS).
"I mean, Jesus H Christ on a motherfucking bike, does this really constitute "news for nerds" ?
What exactly is wrong with IPV4 for Quake ?"
What is wrong with it is the fact that there aren't nearly enough IPv4 addresses to go around. There's an artificial scarcity of addresses that's creating an economy in selling blocks of numbers, which has to be a pretty ludicrous concept when instead we can just make sure there's more numbers than will ever be needed. There's the perpetuation of stupid systems like forcing dialup users to have dynamic addresses. Eventually we will have to switch over to something with a bigger address space. IPv6 is the only viable alternative at the moment, so the more people who are playing with it, programming with it, and generally getting their hands dirty, the smoother the switch-over will be.
"No less than William Gates III (who I normally diagree with vehemently) is on the record as saying 'there is no evidence of consumer demand for an IPv6 offering'"
William Gates III also said that 640k would be enough for everyone.
Of course there's no consumer demand for an IPv6 offering. There's no consumer demand for IPv4 either. Just ask the average consumer if they want IPv4, and they'll say "huh?" There's consumer demand for The Internet, but the consumers don't know, don't care, and don't even think how their packets get from A to B. Thus, it's up to us, the people who actually make this sucker run, to do the dirty work of making sure it will still be running ten years from now.
Charles Miller --
--
The more I learn about the Internet, the more amazed I am
that it works at all.
Re:But IPv4 is more profitable!
by
Wesley+Felter
·
· Score: 2
They will probably have to. Right now, ISPs pay ICANN (through some intermediaries) for IPv4 addresses, so they pass those costs on to customers. I predict that IPv6 addresses will be so cheap that some ISPs will start giving their customers as many as they want, and the ISPs that try to ration v6 addresses will lose business.
IPv6 won't come into real widespread use until a large number of machines can take advantage of it, which given the current OS marketshare means when Windows will support it. Fortunately, there is good news on that front. You can download the beta versions for NT/2000 from MS right now, but around the time Whistler is released, the STABLE release of IPv6 for all Win32 clients should be out and about... once there are lots of consumers that can take advantage of it, I think you will see many more people offering it..... of course, upgrading all those switches and routers is going to be a major PITA. -
The IHA Forums
Slashdot Mirror
Be more specific Taco man. You want Natalie Portman to be the genius that combines those two technologies, or at least brings them to your bedroom. :)
--
I'm not impressed. Give my IPv6 FIRST, then get me the games that can support it. Don't be putting the cart before the horse people. Until the ISP's are supporting IPv6, no one is going to care about this stuff. It's a fun project sure, but it's not even being used on an up-to-date game!
Damn, the cup's half empty!
>but it's not even being used on an up-to-date game!
This is probably because the full source is available for Quake1.
There is much cruelty in the universe, John.
Yeah, we seem to have the tour map.
One of the addresses quoted is:
:-)
3ffe:2b00:100:107::1 (quake6.prav.unisinos.br)
I have no clue - could someone explain what happens when I do a lookup on that hostname and get 200.132.73.102 (on my horrid little IPv4 DNS server...)
And I just don't think hex sextuplets look as good - though they sound better
But getting you IPv6 is more likely to happen when more software supports it. And software is more likely to support it as it becomes more likely that IPv6 will be accepted. In short, it's a chicken-and-egg problem. Anything people can do try drum up support, acceptance, and publicity for IPv6 is going to help. There's no magic switch that can be flipped to make everything IPv6 overnight.
Everyone is saying "We must have ipv6 before we should worry about games". The truth is that you can set yourself up with ipv6 right now! So stop talking and start doing. you can get directions here.
I don't have to have ported a game to IPv6 to be unimpressed. That's the great part about opinions. As for getting it myself, I'm talking about a non-tunneled version of it. Until an ISP hands me an IPv6 addy, I'm not interested. There's no REASON for me to "do it myself" yet.
Several things:
:)
That version of Quake was put up by Viagenie a long time ago. I'm assuming the Slashdotters are calling it "new" just because it was added to the freenet6.net webpage with "New!" written next to it. ("New" refers to the LINK being new, not the game, or even the port!)
Secondly, in order to port Quake to IPv6, they needed Quake source. I'm sure if they'd had access to Quake 3 source at the time, they would have used that. But most game companies don't give away the source to their current money-makers, for some odd reason.
Third, while ISPs dragging their feet getting IPv6 to their customers IS getting annoying, you don't have to wait for them; that's what services like freenet6.net are for. There'll be lag problems due to the tunnelling, but you'll at least have something to play with.
And as for putting the cart before the horse - no one is going to switch to IPv6 if there aren't any programs that take advantage of it. If you try to get IPv6 everywhere before anything uses it, it won't happen. Yes, IPv6 is a superior protocol. That doesn't mean anything. Its human nature (not to mention SOP at ISPs) to bailing-wire and band-aid things to the point of destruction before going through the "work" of an upgrade.
Thankfully, I have native IPv6 access between my workstation, my machines at home, and Ca*Net3, so I'm ready to go.
--
Brandon Hume
hume -> BOFH.Halifax.NS.Ca, http://WWW.BOFH.Halifax.NS.Ca/
Brandon Hume
hume -> BOFH.Halifax.NS.Ca, http://WWW.BOFH.Halifax.NS.Ca/
I have dual redundant tunnels to cisco and Sprintlink via a fractional 10Mb ATM DS-3. Feel free to drop me an e-mail if you are interested. I'd be more then happy to help you get connected. (- the spam-me)
-Pat
NetBSD has had this in its package collection for a long time now, it's available at ftp://ftp.netbsd.org/pub/NetBSD/packages/pkgsrc/ga mes/quake6/README.html.
I do beleive that IPv6 will be supported in Whistler. Maybe after that, we'll begin to see some movement.
IPv6 Quake offers a (partial) solution to this. If it raises the whispers for IPv6 to at least a rumbling, it'll have done amazing good for the Internet as a whole.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
For reasonably fast networks such as a 100MB LAN, most of the latency comes from the software, i.e. the time it takes the operating system to get the data from an application and out onto the network card.
See Tanenbaum's "Computer Networks", 3rd edition, if you are interested in learning more, its a good introduction to networking in general. From pg 562: "Rule #1, CPU speed is more important than network speed". "Long experience has shown that in nearly all networks, operating system and protocol overhead (emphasis mine) dominates actual time on the wire." He goes on to demonstrate that although the theoretical minimum delay getting a frame out onto an 802.3 is 102 microsecs, in practice, it can take up to 15 times longer than that just to get the frame out - the delay is in the software, and protocols make up a large part of the problem.
It isn't about bandwidth. Networks are measured by two things: bandwidth and latency, or the product of these, "bandwidth latency product". For a network to perform well both of these things needs to be good, but on most current networks the bottleneck is not the bandwidth (especially for a game like Quake, where there are lots of small packets, so improving the latency improves the feel of the game more readily than bandwidth improvements.)
Its not uncommon for a packet to be copied in memory 4 or more times from when the application sends it until it gets out onto the wire. Quake uses UDP, which sits on IP, which gets wrapped into an 802.3 frame on a LAN (after doing some arp cache checking of course), and as well as being wrapped up several times, extra copying is also involved in passing the data from the app down to the drivers (and normally within an application itself). All this crap takes time (see "Rule #4: Minimize copying"). Optimising the protocols and software can have a very noticeable effect on network performance. Windows98SE's TCP/IP stack is noticably slower than NT's (I've done some tests copying large numbers of files, typically 30 to 50 % quicker) - this is why they used NT's TCP stack in WinMe, and its also why John Carmack was itching to work on the Linux IP stack a few months ago, to try prod MS into doing something about the lousy performance of Win98's networking. "Rule #5: YOu can buy more bandwidth but not lower delay".
The design and optimization of protocols has a big influence on network performance.
Lots of internet Quake is played at universities, and traffic between Internet2 schools already passes over I2 backbones instead of the "regular" internet. For instance, anybody with a hookup to the Abilene backbone will get great pings to my Q3F server. Unfortunately, your bedroom is probably not considered an academic institution, so you might have a tough time convincing the I2 people to connect you in, IPv6 or not!
Quake1 is GPL, if i remember correctly. Of course, they'll still be wanting to actually make some money off Quake3, but they'll probably open that up in a few years, when Carmack's newest creations are making Q3 look old.
Lag is caused by three things:
a) Einstein (well actually maxwell's laws) i.e. speed of light delay. But this is usually quite small e.g. UK to East Coast US is about 3000km i.e. maybe 30 milliseconds delay allowing for the refractive index of fiber
b) Hops (each machine adds a certain minimum delay, depending on the hardware/software on that host)
c) heavy traffic (when there are lots of packets, the lengths of queues increase, and introduce delays)
Anyway the point is that you can't do anything about a), [short of digging a straight line between the two sites(!), or moving yourselves closer together], b) you might change with extreme difficulty if at all.
However c) is a variable delay and relates to the lack of quality of service guarantees by the network. The kicker here is that the IP6 has protocols that can give guarantees (excepting network failures once the connection is set up).
c) is probably dominant or important in quite a lot of situations, and yes, IP6 does or atleast can help.
Also, in a Quake scenario, any reduction in variation translates into a big help. Variable delays can make it very difficult to play. Playing Quake with what amounts to a higher priority link compared to web page downloads or ftp sessions is a godsend - it isn't as if Quake uses much bandwidth, but it is very timing sensitive.
-WolfWithoutAClause
"Gravity is only a theory, not a fact!"The average ISP runs almost all Cisco kit - even IOS 12.1(5)T, one of the most recent ones, does not support IPv6, and most ISPs will be running earlier versions. I believe IPv6 support from Cisco (in non-beta IOS versions) will be out this year. I believe Bay and many other routers already support IPv6, but that doesn't matter for most ISPs.
There's also a lot of planning required in how exactly the ISP should switch on IPv6 in their networks - their whole routing setup and DNS infrastructure needs extending, for a start, not to mention the various 6-to-4 conversion schemes. It's not just a matter of turning IPv6 on.
When IPv6 is enabled, you should not have to type 128 bit addresses in any case (did you type in the address for slashdot.org, or did you use DNS).
I don't understand why people are still using 6bone tunnels when 6to4 is easier and more efficient.
What is wrong with it is the fact that there aren't nearly enough IPv4 addresses to go around. There's an artificial scarcity of addresses that's creating an economy in selling blocks of numbers, which has to be a pretty ludicrous concept when instead we can just make sure there's more numbers than will ever be needed. There's the perpetuation of stupid systems like forcing dialup users to have dynamic addresses. Eventually we will have to switch over to something with a bigger address space. IPv6 is the only viable alternative at the moment, so the more people who are playing with it, programming with it, and generally getting their hands dirty, the smoother the switch-over will be.
William Gates III also said that 640k would be enough for everyone.
Of course there's no consumer demand for an IPv6 offering. There's no consumer demand for IPv4 either. Just ask the average consumer if they want IPv4, and they'll say "huh?" There's consumer demand for The Internet, but the consumers don't know, don't care, and don't even think how their packets get from A to B. Thus, it's up to us, the people who actually make this sucker run, to do the dirty work of making sure it will still be running ten years from now.
Charles Miller
--
The more I learn about the Internet, the more amazed I am that it works at all.
They will probably have to. Right now, ISPs pay ICANN (through some intermediaries) for IPv4 addresses, so they pass those costs on to customers. I predict that IPv6 addresses will be so cheap that some ISPs will start giving their customers as many as they want, and the ISPs that try to ration v6 addresses will lose business.
The bandwidth is useless if the content is lacking.
IPv6 won't come into real widespread use until a large number of machines can take advantage of it, which given the current OS marketshare means when Windows will support it. Fortunately, there is good news on that front. You can download the beta versions for NT/2000 from MS right now, but around the time Whistler is released, the STABLE release of IPv6 for all Win32 clients should be out and about... once there are lots of consumers that can take advantage of it, I think you will see many more people offering it..... of course, upgrading all those switches and routers is going to be a major PITA.
-
The IHA Forums
Natural != (nontoxic || beneficial)