From: craigberry@nospam.mac.com   
      
   On 10/11/25 6:35 AM, Dan Cross wrote:   
   > In article <10cauq9$20cl$1@dont-email.me>,   
   > Simon Clubley   wrote:   
   >> On 2025-10-10, Dan Cross  wrote:   
   >>> In article <10c89re$2n78q$1@dont-email.me>,   
   >>> Simon Clubley   wrote:   
   >>>> On 2025-10-08, Arne Vajhøj  wrote:   
   >>>>> VSI has published final agenda:   
   >>>>>       https://events.vmssoftware.com/bootcamp-2025/agenda   
   >>>>>   
   >>>>   
   >>>> That's a very interesting range of topics. It's starting to feel   
   >>>> like the old DECUS conferences of yesteryear.   
   >>>>   
   >>>> One thing of special interest is that the rewrite of the TCP/IP stack   
   >>>> is back on the WIP list. This was tried in the past by VSI and then   
   >>>> dropped. Interesting that the FreeBSD stack was chosen as the starting   
   >>>> point for this.   
   >>>   
   >>> I'm curious as to your thoughts on the FreeBSD stack.  My sense   
   >>> is that it is one of the better implementations out there, and   
   >>> has had a ton of work done on it (including recently).  It also   
   >>> has a permissive license that makes it suitable for integration   
   >>> into a proprietary system like VMS.   
   >>>   
   >>   
   >> That's exactly my point. It's an excellent starting point for   
   >> a VMS-specific TCP/IP stack.   
   >>   
   >> It's vastly better than the original UCX disaster area and looks to   
   >> be far better than whatever they used for the UCX 5 (TCP/IP Services)   
   >> rewrite.   
   >   
   > Agreed.  Where does Process Software fit in all of this?   
   > Multinet always seemed very good.   
      
   That's the thing Simon mentioned that had been dropped. VSI announced a   
   new TCP/IP stack based on Multinet but later withdrew it.  As far as I   
   know, it's still available independently from Process Software.   
      
   I believe the existing DEC/Compaq/HP(E)/VSI stack is based on BSD, but   
   BSD from 25+ years ago, so it does make sense to update to current,   
   which after that much time is probably a complete rewrite.   
      
   Hopefully the Bootcamp talk will explain what exactly is getting redone.   
     Is it only the socket libraries, and the layer that does QIOs to the   
   LAN driver and the management interfaces will stay the same?  Or are   
   they revisiting VCI 2.0, an update to the LAN driver that was planned   
   and canceled during the HP(E) era and would have overhauled how the TCP/   
   IP stack interfaces with networking hardware (and without which,   
   networking speed on VMS will probably never be competitive)?   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)