Re: question

huston@ps73.ako.dec.com Wed, 27 January 1993 17:27 UTC

Message-Id: <9301271723.AA09563@ps73.ako.dec.com>
To: Frank T Solensky <solensky@andr.ub.com>
Cc: criteria@ftp.com
Subject: Re: question
In-Reply-To: Your message of "Wed, 27 Jan 93 10:40:37 EST." <9301271540.AA07688@fenway.andr.UB.com>
Date: Wed, 27 Jan 1993 12:23:02 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: huston@ps73.ako.dec.com

>>What portion of the total cost of modifying host (either for SIP or for TUBA)
>>would be in the items above the Network Layer (e.g. applications) ?
>
>Again, gut feel (and before my first cup of coffee): not a large portion, and
>probably not much different between the proposals.

Probably not that much different between the proposals I would agree with...

>apps themselves rarely deal with anything more involved that the socket
>number.

I believe that a lot of applications do things like:

	int ip_addr;

	ip_addr = ntohl(*(int *)ptr);
	printf("%d.%d.%d.%d", ip_addr & 0xFF000000 >> 24,
			      ip_addr & 0x00FF0000 >> 16,
			      ip_addr & 0x0000FF00 >> 8,
			      ip_addr & 0x000000FF        );

etc.

So, while it may not be technically very difficult, it may be tough to weed
out all the assumptions built on the IPv4 address format.


Steve Huston
huston@ps73.ako.dec.com
+1 508 264 7117
On contract to, but not speaking for, Digital Equipment Corporation

question Eric Fleischman
Re: question Fred Baker
Re: question Tony Li
Re: question William Allen Simpson
Re: question Frank T Solensky
question yakov
Re: question Frank T Solensky
Re: question huston
Re: question Noel Chiappa
Re: question Scott_Brim
Re: question William Allen Simpson