[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RRG] Is 12 bytes really so scary?



In einer eMail vom 08.12.2007 12:00:49 Westeuropäische Normalzeit schreibt HeinerHummel@aol.com:
In einer eMail vom 08.12.2007 09:35:06 Westeuropäische Normalzeit schreibt rw@firstpr.com.au:
12 bytes is all we need for IPv4.
360 + 180 = 540 = 512 +28, i.e. a number that fits into 10 bits.
slight correction: 360 takes 9 bits, 180 takes 8 bits. I.e. 17 bits altogether.
Hence 17 additional bits in the IPv4 header would enable scalable IPv4-forwarding even if the internet were thousand times bigger and denser.
 
Further side effects:
- 20 times faster next hop determination
 
- IPv4 address depletion becomes a non issue provided that the IRTF  seriously cared about
identifier/location split rather than kept going to sell the "EID/RLOC"-2 loose hops forwarding as a separation of identifier and location. Within the same city there may live several Mr.Smith ! Each of whom
may also roam BTW ! But each would have a different social security card number or passport number which is not affected in case of roaming ! The currently discussed solutions just prove that the longly theoretical loc/id  discussion a year ago was just done to pave the ground for LISP and just stopped all of a sudden.
IPv4 addresses could be just geopatch locally unique (therefore no problem with IPv4 address depletion).
But agreed, roaming to some other geopatch must be enabled, too, i.e. must not wind up in a situation where there are several IP-phones with the same IPv4-address. Will say: identifier/location split is needed,
but is not provided and dealt with by the current LISP (agreed, I haven't yet provided a solution to this problem (in case of roaming) either.But why should I, if no one cared anyway).
 
 
Heiner