Subject: Re: kernel ip_randomid() and libc randomid(3) still "broken"
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Charles M. Hannum <abuse@spamalicious.com>
List: tech-net
Date: 11/17/2003 02:51:04
On Monday 17 November 2003 02:26 am, Jonathan Stone wrote:
> I've heard a couple of people say that Sun has a smarter random IP_ID
> allocation, maybe based on inpcb state?. I haven't tried to implement
> that, as I've never heard how it handles IDs for IP output packets not
> associated with an inpcb (e.g., outbound ICMP, or multicast/broadcast).

As it was explained to me, Solaris keeps track of per-host state, and has a 
unique ID space for each (destination or {source,destination} pair?).  The 
main purpose being to avoid having IDs wrap around quite as frequently with 
high-bandwidth links.