A First Byte of IPv8
To NANOG list members... Thanks for the interesting comments on IPv6 and IPv4. It is always interesting to see why IPv6 is the way it is. As you might know, IPv8 is not that way... All you have to do is look at the first byte of IPv8 to realize that it was designed for performance, not to satisfy or calm the press. There are 8 bits in that byte and they have been carefully allocated as follows: 1OSSSDDD 1 - The IPv8 flag bit (IPv4 and IPv6 are 0) O - Options Present (0 - no, 1 - yes) SSS - Source Galaxy DDD - Destination Galaxy By looking at only this first byte, routers can sort packets destined for one of the 8 Galaxies very quickly. Routers can also discover whether the header contains options and route those packets locally for further processing. IPv8 has exactly the same size header as IPv4. It is therefore easy to learn and understand. IPv8 has been designed with 48 byte ATM payloads in mind. Therefore, the performance people and QOS people will be happy. IPV8 has been designed for the emerging world of circuit switched routing, where packets are moved a long way without "hops" so that they can be delivered to their intended destination as quickly as possible. IPv8 was designed to address the routing problem as well as the Address Management and Governance problems. The 11 bits of addressing added to the 32 bit IPv4 addresses are divided into 3 bits and 8 bits and a natural hierarchy takes shape. The core has 8 regions each with 256 distribution centers and from there full 32 bit Internets. This is 2,048 address spaces. If these address spaces are managed well, they can be used in IPv6 addressing. NANOG people may want to consider focusing on address space management without being concerned about the protocol header format. NANOG could help to make a contribution in making sure these emerging address spaces are managed the way your members desire. That would be a major contribution to the future of all the 2,048 Internets. Jim Fleming Unir Corporation IBC, Tortola, BVI
On Tue, 4 Nov 1997, Jim Fleming wrote:
All you have to do is look at the first byte of IPv8 to realize that it was designed for performance, not to satisfy or calm the press.
Could you possibly supply a URL or other pointers for the RFC or other appropriate specifications for "IPv8"?
IPV8 has been designed for the emerging world of circuit switched routing, where packets are moved a long way without "hops" so that they can be delivered to their intended destination as quickly as possible.
Also, where can I obtain the source code of the reference implementation? There is a reference implementation, isn't there? Any ideas as to when I can deploy Ipv8 on my Cisco?
Brian Tackett writes...
IPV8 has been designed for the emerging world of circuit switched routing, where packets are moved a long way without "hops" so that they can be delivered to their intended destination as quickly as possible.
Also, where can I obtain the source code of the reference implementation? There is a reference implementation, isn't there? Any ideas as to when I can deploy Ipv8 on my Cisco?
I wouldn't want to see any kind of "reference implementation" even be attempted until the protocol itself has been reviewed. That at least gives others the opportunity to suggest changes the original author may have overlooked, and get them into the reference implementation. A lot of people have this perception that a proposal cannot be judged until they see source code. While many protocols have probably been done that way, many others have not and instead get reviews on the design theory before any code is ever written. -- Phil Howard | crash285@s5p6a6m2.org w3x0y9z8@nowhere1.com end2it61@noplace1.net phil | blow6me2@s0p3a1m0.com suck2it3@anywhere.com w8x4y3z7@anywhere.org at | a1b6c8d0@no21ads1.edu die0spam@lame1ads.com end5ads7@anywhere.edu milepost | stop0ads@nowhere0.edu ads6suck@s2p8a0m1.org a5b1c6d1@noplace5.net dot | a8b2c2d5@dumb8ads.com eat97me8@no72ads4.com no8way23@noplace1.net com | end0ads4@s5p4a8m6.edu stop1ads@spam0mer.org stop0ads@spammer9.org
On Tue, Nov 04, 1997 at 12:44:43PM -0600, Phil Howard wrote:
Brian Tackett writes...
IPV8 has been designed for the emerging world of circuit switched routing, where packets are moved a long way without "hops" so that they can be delivered to their intended destination as quickly as possible.
Also, where can I obtain the source code of the reference implementation? There is a reference implementation, isn't there? Any ideas as to when I can deploy Ipv8 on my Cisco?
I wouldn't want to see any kind of "reference implementation" even be attempted until the protocol itself has been reviewed. That at least gives others the opportunity to suggest changes the original author may have overlooked, and get them into the reference implementation.
Call this one a successful troll, Jim. :-) Cheers, -- jr '<CHOMP>' a -- Jay R. Ashworth jra@baylink.com Member of the Technical Staff Unsolicited Commercial Emailers Sued The Suncoast Freenet "Pedantry. It's not just a job, it's an Tampa Bay, Florida adventure." -- someone on AFU +1 813 790 7592
On Tue, Nov 04, 1997 at 10:55:15PM -0600, Jim Fleming wrote:
All you have to do is look at the first byte of IPv8 to realize that it was designed for performance, not to satisfy or calm the press.
Damn; I could'a had a v8. Cheers, -- jr 'you know there's an RFC in here, right?' a -- Jay R. Ashworth jra@baylink.com Member of the Technical Staff Unsolicited Commercial Emailers Sued The Suncoast Freenet "Pedantry. It's not just a job, it's an Tampa Bay, Florida adventure." -- someone on AFU +1 813 790 7592
participants (4)
-
Brian Tackett
-
Jay R. Ashworth
-
Jim Fleming
-
Phil Howard