[members-discuss] Technical solution to resolve the IPv4 Exhaustion problem and to add more 4, 294, 967, 296 IPv4 addresses that are needed in the world

Christian,

Let me explain:

"The ip address is the clients identity. That is how it is routed. Routing cannot route ipv4 and ipv4+ packets to different destinations because packets are routed by the destination address only."  - Routers do not route ip packets based on the 32 bits of the destination address in binary format, they route based on the representation of [0-255].[0-255].[0-255].[0-255] - four decimal numbers in specific ranges between 0 to 255 separated by dots, each decimal number is one byte, in total 4 bytes like they are in the source address or destination address field in the ip header.

If the reserved bit flag in the ip header is set to '1' (the reserved bit is not a new bit to be added to the ip header, it is already exist and it is currently set to '0' , we will set it to '1' ) - then the exact same four bytes of the source address (just for the example) will be represented as [0-65535].[0-65535] , meaning two decimal numbers with bigger ranges separated by one dot (the ranges are bigger because each of these two decimal numbers if of two bytes - in total four bytes just like [0-255].[0-255].[0-255].[0-255] ) - by setting the reserved bit to '1' or to '0' we can tell the router if the four bytes - meaning the 32 bits of the source address (just for the example) are in the format of [0-255].[0-255].[0-255].[0-255] if reserved bit is '0' or in the format of [0-65535].[0-65535] if reserved bit '1' , based on it the router will know what is the ip address and then will route it accordingly (routers that have no more than two physical routes - will not need to check the reserved bit in order to know what is the format, they will just forward the ip packet to the other physical route).


Regarding:
"The string represenation of the address is not the criteria how existing socket api work.  The api work internally with 32 bit values.  There is no formatting in them." - It is related to the implementation of the operating system internally in the operating system, single patch to the operating system through the automatic update mechanism of the operating system - will resolve it (As part of global IPv4+ implementation there will be a single remote update to each operating system that will want to support IPv4+).

Regarding:
"You can only have more addresses if you add bits to the addresses." - Please see my first explanation above on how the same four bytes can be used in ip packets with different representation by using the currently unused reserved bit flag in the ip packet.


Respectfully,
Elad
________________________________
From: Christian Kratzer <ck at cksoft.de>
Sent: Sunday, April 26, 2020 11:31 AM
To: Elad Cohen <elad at netstyle.io>
Cc: Tobias Lehner <tl at hartl-edv.de>; 'noc' <noc at xervers.pt>; Ed Campbell <campbell at inca.ie>; members-discuss at ripe.net <members-discuss at ripe.net>
Subject: Re: [members-discuss] Technical solution to resolve the IPv4 Exhaustion problem and to add more 4, 294, 967, 296 IPv4 addresses that are needed in the world

Hi Elad,

On Sun, 26 Apr 2020, Elad Cohen wrote:
> Christian,
>
> I'm sorry to write, but you didn't understand how IPv4+ works. And everything that you wrote regarding IPv4+ is completely incorrect.

Everything about IPv4+ is completely inpractical.

> "There is no connectivity between IPv4 and IPv4+" - IPv4+ is IPv4, exact same protocol.
>
> "you would need routers to support 33 bit routes which is not going to happen" - This is completely incorrect, route bits are exactly the same.

The ip address is the clients identity. That is how it is routed. Routing cannot route ipv4 and ipv4+ packets to different destinations because packets are routed by the destination address only.

So while the core networks would transpart packets between LIR the LIR themselves would not be able to route packets to different clients.

> "To enable a client to connect to both the IPv4 and IPv4+ internet it seems to me that you would need at least another address family in the socket protocols which is also a massive overhead.  The formatting of the address as two 16 bit values instead of four 8 bit values does not fix the issue in the clients ipv4 stack." - No another address family is needed, the source address and destination are exactly in the four bytes each as they are now - the only difference is the application layer in the operating system - based if the single reserved bit flag is on or off - the ip address will be displayed with one dot (IPv4+) or with three dots (IPv4).

The string represenation of the address is not the criteria how existing socket api work.  The api work internally with 32 bit values.  There is no formatting in them.

> "You cannot route ipv4 and ipv4+ in the same global internet if they are two seprate networks and if the addresses mean different things depending on arbitrary address bits." - It is the same network, IPv4 (I'm calling it IPv4+ to represent the one-dot addressing to higher application layers, but it is the same IPv4 packets, same IPv4 network)

You can only have more addresses if you add bits to the addresses.

> I'm not against IPv6, IPv6 and IPv4 will always co-exist in some way, IPv4+ brings more ip addresses to IPv4, it doesn't disturb a bit IPv6.

it breaks ipv4.

Greetings
Christian

--
Christian Kratzer                   CK Software GmbH
Email:   ck at cksoft.de               Wildberger Weg 24/2
Phone:   +49 7032 893 997 - 0       D-71126 Gaeufelden
Fax:     +49 7032 893 997 - 9       HRB 245288, Amtsgericht Stuttgart
Mobile:  +49 171 1947 843           Geschaeftsfuehrer: Christian Kratzer
Web:     http://www.cksoft.de/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.ripe.net/ripe/mail/archives/members-discuss/attachments/20200426/9a3a9757/attachment.html>