[Info-vax] VMS Software Inc. OpenVMS 8.4-1H1 Boots on i4 System

[Stephen Hoffman]

On 2015-03-20 04:31:10 +0000, David Froble said:

> Bill Gunshannon wrote:
> 
>> I know how you feel.  This past week I put about 5 bookshelf feet of
>> Windows Manuals in the recycle bin.  Somehow, I don't think NT, 2000,
>> 98 or even XP are coming back.
>> 
>> And today Iearned Server 2003 is EOS in July.  Not surprising as I have
>> already noticed some of the stuff they are doing to force the issue.
>> I guess they learned their lesson with XP.  If you don;t force people
>> off it, they will stay.
> 
> I've been heard to mutter "just because it's old doesn't mean it won't 
> continue to do what it's always done".
> 
> Software doesn't wear out.

There's an OpenSSL security update announced this week, including 0.9.8af.

There's a new vulnerability in the disclosure process, too.  This new 
vulnerability is reportedly more severe than the disclosure this week, 
too.

There's also that the OpenSSL folks are retiring the rather archaic 
branch that the HP OpenSSL port is based on.  Security patches for 
0.9.8 are ending in 2015; at the end of this year.

What HP and VSI have in mind here and what their OpenSSL patch 
schedules and update schedules might be, I don't know.

I've been testing some code with the current branch, and there have 
been one or two very minor changes required to the application code 
from what was needed with the archaic branch.

> HW does, and when old software cannot use new HW, that can be an issue.

True.  Sometimes old software can get slow at scale, even on older 
hardware.   The volume of data that more than a few programs are 
dealing with tends to increase, too.  For a specific instance of this 
slowdown at scale, linked lists are used by more than a few application 
programs, and are O(n) for insertions and deletions.  As configurations 
get bigger, the existing and working linked list code doesn't scale.  
<http://en.wikibooks.org/wiki/Data_Structures/Tradeoffs>

> I guess that's why I see some value in emulators and VM stuff, even if 
> I haven't done anything with any of that yet.

Emulators can be the best of some rather bad choices.

> The boss has been talking to David Turner lately.  I created a monster. 
>   Now he's asking me if we wouldn't want to consider Alpha emulators 
> instead of itanics.  I of course don't have a clue.

Any of the recent Itanium boxes are faster than the Alpha emulators, 
and many of the Itanium boxes are unfortunately also usually louder.  
The Itanium boxes have the advantage of being more familiar to many 
folks when you need help — as you're running a configuration that other 
folks usually already have, and that VSI probably directly supports — 
and the hardware configurations are generally simpler than the 
emulators — you're not dealing with the underlying host system and the 
emulator software and the emulator console and the emulator virtual 
networking support.  The emulators are definitely getting better about 
their respective virtual networking, but it's still somewhat of a 
quagmire with some.  Not that EFI, the BMC and the iLO MP are shining 
examples of user interface design clarity, of course.


-- 
Pure Personal Opinion | HoffmanLabs LLC