Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On Sat, 27 Apr 2002 [EMAIL PROTECTED] wrote: So if your P2P application is IPv6 compatible, you can get a semi permanent IPv6 IP automatically from a server, and thereafter do peer to peer, just as if you were full, no kidding, on the internet. This nicely solves the problem with NATs, true. However, most firewalls I know are there for security reasons. Those will likely be adapted to work for 6to4 as well. The transition period will likely see some cracks where p2p can work, but I suspect those will be closed in due course. Sampo Syreeni, aka decoy - mailto:[EMAIL PROTECTED], tel:+358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2
Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On Sun, 28 Apr 2002, Lucky Green wrote: I concur. In fact, I was surprised that not a single one of the many P2P solutions presented at the recent excellent CODECON made any mention of support for IPv6, which can be easily be added to just about any P2P application, while every presenter bemoaned the fact that the existence of NAT's between nodes provided severe design and functionality constraints. The fix is not just obvious, it is downright trivial on any modern OS. Stop playing with toys. Plan 9. -- The law is applied philosophy and a philosphical system is only as valid as its first principles. James Patrick Kelly - Wildlife [EMAIL PROTECTED] www.ssz.com [EMAIL PROTECTED] www.open-forge.org
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On 28 Apr 2002 at 16:20, Morlock Elloi wrote: How exactly does the introduction of IPV6 on a machine that is NAT-ted by the ISP who doesn't give shit about IPV6 help the situation ? James A. Donald: To connect to the IPV6 world from inside a NAT network, you need a machine that is both inside and outside the NAT network, a gateway machine that has an IP4 an external address, even if only a dynamic address. Then all machines on the inside can talk to the outside through that machine, thus they can all receive quasi static IP6 addresses, even though not even the gateway machine possesses a static IP4 address. To clarify, this means that if you have a home network with a gateway computer, you can probably get global static IPV6 addresses for all the machines of your home network, though you might have trouble getting software to use this, or finding people to who can access your computers in IPV6 However, for a corporation, such measures make sense and are useful, because it means they can videoconference within the corporation, and also with other corporations that have adopted the same measure -- video conferencing being the P2P app used by people who are willing to pay money for it.
Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On Sat, 27 Apr 2002 [EMAIL PROTECTED] wrote: So if your P2P application is IPv6 compatible, you can get a semi permanent IPv6 IP automatically from a server, and thereafter do peer to peer, just as if you were full, no kidding, on the internet. This nicely solves the problem with NATs, true. However, most firewalls I know are there for security reasons. Those will likely be adapted to work for 6to4 as well. The transition period will likely see some cracks where p2p can work, but I suspect those will be closed in due course. Sampo Syreeni, aka decoy - mailto:[EMAIL PROTECTED], tel:+358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2
Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
-- On 29 Apr 2002 at 14:58, Sampo Syreeni wrote: [IPv6] nicely solves the problem with NATs, true. However, most firewalls I know are there for security reasons. Those will likely be adapted to work for 6to4 as well. The transition period will likely see some cracks where p2p can work, but I suspect those will be closed in due course. Customers want P2P. Businesses will supply it. The reason they are not supplying it now is that there is an IP shortage. --digsig James A. Donald 6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG LFpwFHMRzVb3bItJnefOQKed+0h+Ra8Z4V5mtA1b 4U8h947/ql0vOFSk9s9IMkJ1fW8pZPVSSfyvCOL0R
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
James wrote: IPV6 to the rescue. Every network behind a NAT router should set up a 6to4 tunnel, probably some time early in 2003. IPv6 is almost source code compatible with IPv4, so every application should soon be recompiled to be IPv6 compatible. Every computer with a recent operating system, for example recent linux kernels and windows XP, theoretically supports IPv6. So if your P2P application is IPv6 compatible, you can get a semi permanent IPv6 IP automatically from a server, and thereafter do peer to peer, just as if you were full, no kidding, on the internet. I concur. In fact, I was surprised that not a single one of the many P2P solutions presented at the recent excellent CODECON made any mention of support for IPv6, which can be easily be added to just about any P2P application, while every presenter bemoaned the fact that the existence of NAT's between nodes provided severe design and functionality constraints. The fix is not just obvious, it is downright trivial on any modern OS. --Lucky
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On 28 Apr 2002 at 0:15, Lucky Green wrote: I concur. In fact, I was surprised that not a single one of the many P2P solutions presented at the recent excellent CODECON made any mention of support for IPv6, which can be easily be added to just about any P2P application, while every presenter bemoaned the fact that the existence of NAT's between nodes provided severe design and functionality constraints. The fix is not just obvious, it is downright trivial on any modern OS. --Lucky As I recall, the reptile guy said the reason he was using the jdk 1.4 beta was its support for IPv6. George
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
I concur. In fact, I was surprised that not a single one of the many P2P solutions presented at the recent excellent CODECON made any mention of support for IPv6, which can be easily be added to just about any P2P application, while every presenter bemoaned the fact that the existence of NAT's between nodes provided severe design and functionality constraints. The fix is not just obvious, it is downright trivial on any modern OS. How exactly does the introduction of IPV6 on a machine that is NAT-ted by the ISP who doesn't give shit about IPV6 help the situation ? Are you saying that putting *any* software/OS on a computer behind NAT/firewall somehow magically gives it visibility from the outside ? = end (of original message) Y-a*h*o-o (yes, they scan for this) spam follows: Yahoo! Health - your guide to health and wellness http://health.yahoo.com
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
-- On 28 Apr 2002 at 16:20, Morlock Elloi wrote: How exactly does the introduction of IPV6 on a machine that is NAT-ted by the ISP who doesn't give shit about IPV6 help the situation ? To connect to the IPV6 world from inside a NAT network, you need a machine that is both inside and outside the NAT network, a gateway machine that has an IP4 an external address, even if only a dynamic address. Then all machines on the inside can talk to the outside through that machine, thus they can all receive quasi static IP6 addresses, even though not even the gateway machine possesses a static IP4 address. The question then is when will ISP's feel pressure to provide such a a gateway? --digsig James A. Donald 6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG u6gm2uaf41VUVwgcdHrLWjfpoumqf3gh0alLqCQA 4twho9x1bOXnA+ZB85c2gi3TMua3r+rWLXHEnVNgy
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On 28 Apr 2002 at 16:20, Morlock Elloi wrote: How exactly does the introduction of IPV6 on a machine that is NAT-ted by the ISP who doesn't give shit about IPV6 help the situation ? James A. Donald: To connect to the IPV6 world from inside a NAT network, you need a machine that is both inside and outside the NAT network, a gateway machine that has an IP4 an external address, even if only a dynamic address. Then all machines on the inside can talk to the outside through that machine, thus they can all receive quasi static IP6 addresses, even though not even the gateway machine possesses a static IP4 address. To clarify, this means that if you have a home network with a gateway computer, you can probably get global static IPV6 addresses for all the machines of your home network, though you might have trouble getting software to use this, or finding people to who can access your computers in IPV6 However, for a corporation, such measures make sense and are useful, because it means they can videoconference within the corporation, and also with other corporations that have adopted the same measure -- video conferencing being the P2P app used by people who are willing to pay money for it.
RE: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
-- On 28 Apr 2002 at 16:20, Morlock Elloi wrote: How exactly does the introduction of IPV6 on a machine that is NAT-ted by the ISP who doesn't give shit about IPV6 help the situation ? To connect to the IPV6 world from inside a NAT network, you need a machine that is both inside and outside the NAT network, a gateway machine that has an IP4 an external address, even if only a dynamic address. Then all machines on the inside can talk to the outside through that machine, thus they can all receive quasi static IP6 addresses, even though not even the gateway machine possesses a static IP4 address. The question then is when will ISP's feel pressure to provide such a a gateway? --digsig James A. Donald 6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG u6gm2uaf41VUVwgcdHrLWjfpoumqf3gh0alLqCQA 4twho9x1bOXnA+ZB85c2gi3TMua3r+rWLXHEnVNgy
Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
-- On 18 Feb 2002 at 14:37, Sampo Syreeni wrote: we still need one of the machines to be outside a firewall. I think what anonymous is describing is the situation when each and every non-corporate customer is behind a firewall owned by an ISP, corporations shield their employees behind one of their own, where making a profit on non-firewalled connectivity and/or proxies opens one up to expensive lawsuits and where non-firewalled connectivity is too expensive to be widely bought just for non-profit use. A balkanized Internet -- it's a bleak prospect, and the signs are that's what we're slowly sliding towards, at least at the moment. IPV6 to the rescue. Every network behind a NAT router should set up a 6to4 tunnel, probably some time early in 2003. IPv6 is almost source code compatible with IPv4, so every application should soon be recompiled to be IPv6 compatible. Every computer with a recent operating system, for example recent linux kernels and windows XP, theoretically supports IPv6. So if your P2P application is IPv6 compatible, you can get a semi permanent IPv6 IP automatically from a server, and thereafter do peer to peer, just as if you were full, no kidding, on the internet. IPv6 is being driven largely by demands for peer to peer. Phone companies want to give each cell phone an IP address (being done already in Japan). Businesses want to videoconference. --digsig James A. Donald 6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG bZKSBSjNMZ8oyTPKxYPWV6RgQaZwt6mM8R/eZcyC 4JMoDxw0Bk4JN3l44B4rbsaulwD+bcAkYzr9R3Vs6
p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
I think the asymmetric up/down speed is not as much a problem for peer2peer as anonymous fears. Morpheus has demonstrated that the approach of having a single request served by multiple servers works well. A cable modem users download speed can be merrily supplied by dozens of even dialup, or other cable modem users thin pipe uploads speeds. Morpheus seems to be able to tunnell through even to corporate firewalls with the approach (I presume) that the firewalled / unreachable host maintains a connection to the super-node, and when someone wants to connect to it and can't they connect to the super-node and the super-node tells the unreachable node over the already open connection to connect back to the connecting machine. Of course this can't work (without moving data via the super-node) between two unreachable machines, but the balance seems to be sufficiently in favor of reacable machines that I don't see it currently presents a problem. Adam Anonymous writes: Few impressions after just closed CodeCon 2002 (http://codecon.org) - NATing is successfully choking P2P. All solutions require subpoenable and destroyable proxies. Address space is owned by whoever owns the network. If you have no address you can't publish. - It is to be expected that ISPs will further limit upload bandwidth. Even 50:1 download/upload max bandwidth ratio will not affect bona fide commercial apps and will fuck all P2P big time. This is trivial to do. Upload will become more and more expensive and in some juristictions subject to licensing (WHY DO YOU NEED 20 TEDDY BEARS?)
Re: p2p and asymmetric bandwidth (Re: Fear and Futility at CodeCon)
On Mon, 18 Feb 2002, Adam Back wrote: and when someone wants to connect to it and can't they connect to the super-node and the super-node tells the unreachable node over the already open connection to connect back to the connecting machine. Of course, that approach could be extended do the point where there is no essential difference between a proxy (here, the supernode) and a usual client. Of course this can't work (without moving data via the super-node) between two unreachable machines, but the balance seems to be sufficiently in favor of reacable machines that I don't see it currently presents a problem. Yes, we still need one of the machines to be outside a firewall. I think what anonymous is describing is the situation when each and every non-corporate customer is behind a firewall owned by an ISP, corporations shield their employees behind one of their own, where making a profit on non-firewalled connectivity and/or proxies opens one up to expensive lawsuits and where non-firewalled connectivity is too expensive to be widely bought just for non-profit use. A balkanized Internet -- it's a bleak prospect, and the signs are that's what we're slowly sliding towards, at least at the moment. Sampo Syreeni, aka decoy - mailto:[EMAIL PROTECTED], tel:+358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2