[Info-vax] Bare Metal VMS (Frame.Work Laptops) <<<< complete answer late answer was incomplete, sorry

[VMSgenerations]

Le 12/11/2024 à 23:59, Subcommandante XDelta a écrit :
> Just some blue sky kite flying, perhaps not grounded on the current
> revision levels of VSI reality - to be honest, I don't diligently keep
> up with the technicalities.
> 
> But wouldn't be lovely if we had laptops that were running bare metal
> VMS? - the likes of which we haven't seen since the Alpha Tadpole
> (IIRC).
> 
> I know "bare metal VMS" sounds like some unsafe sexual practice - and
> I don't know the ins and outs (so to speak) of why it isn't common
> practice, apart from the issues of device drivers writing, given the
> plethora of devices and modules, that might have to be catered for.
> 
> Which is why the Frame.Work laptop initiative is a bright blessing:
> 
> https://frame.work/
> 
> I think Big Ken would have had grudging respect, or, indeed,
> unreserved respect for what they have achieved.
> 
> If I wasn't such a "TrackPointer Tragic" ThinkPad veteran, this would
> be my laptop platform of choice - CRU (FRU) heaven, and the
> anthithesis of "enshittifcation".
> 
> As it is my current generations of ThinkPads will see me out, keeping
> a weather eye on my running down system clock.
> 
> Quite possibly only a small set of device drivers would need to be
> considered, if bare metal VMS was offered on their hardware platforms,
> and given the elegance of their construction, field service
> engineering would not have to be considered - people can roll their
> own - nothing is soldered down that strictly doesn't have to be - CRU
> heaven.
> 
> I think it would be the perfect marriage of VMS with third-party
> hardware, and it would be a boon to the VMS ecosystem, to have such
> available - it might even encourage new VMS shops being established -
> new clients for VSI.
> 
> Given the modular I/O port design, they would be a boon to VSI
> engineering for research, VSI Quality Assurnce for testing, VSI sales
> for demonstrations, and existing, and possibly new, customers for
> evaluating.
> 
> You could have two Ethernet port modules one for networking the other
> for clustering computer interconnect, and two USB ports for Disk
> shadow sets - the laptops would be cheap, perfect, Lego blocks for
> exploring the VMS clustering and Shadow Set chops of VMS.
> 
> Further VSI could liaise with Frame.Work, and commission them to
> design a Server Blade board, for running bare metal VMS, an Intel one,
> and an AMD one, utilising "Gruntmaster 6000" (cf Dilbert, circa 2000)
> CPUs and "GruntMaster 6000" GPUs (which seen all the rage these days,
> with the AI hype).
> 
> The server board would have a RS232 port (well, just because!), but it
> would also have (say) four narrow expansion slots to cater for DEC
> specific I/O considerations - another boon for the VMS ecosystem.
> 
> And one of the genius aspects of the Frame.Work way, is that like the
> Ship of Theseus, individual components, can be easily upgraded.
> 
> VSI management might have kittens and conniptions about how to price
> VSI/VMS licenses for the laptops - but what is not made in margins,
> may well be made in volume - I am sure if such laptops were available
> to the VMS ecosystem, that they would fly off the shelves.
> 
> The 13" laptops for the Office, and portability, and the 16" laptops
> for scientific and engineering workstations.
> 
> Anyways, that's my back of a beer coaster (lagered products) thesis &
> proposal.
> 
> I'll leave it up to those who know what they are talking about to
> weigh knowledgeably in.
> 
> And what might the VMSGenerations group think of it all?
Dear Subcommandante Xdelta,


We really appreciate your interest in our work in general, and we're 
glad you're asking us for advice.

The question you are asking is particularly interesting. It overlaps 
with the concerns of our group but not in exactly the same terms.

The users and professionals we represent essentially use VMS in a server 
environment most often linked to high availability services, quality 
requirements and long or very longterm durability. Remote management is 
also a common request. So, at this moment, the world of laptops is not 
part of our members’ priorities and concerns.

You quote Ken Olsen and we understand that your proposal is a vision. 
Linking it to the frame.work operation is really relevant, since it's 
part of the Digital culture to provide innovative alternatives, based on 
technical issues that may pose a problem. With the introduction of 
minicomputers, PdP and VAX, Digital transformed the use of computers in 
the era of « mainstream of main frames ». In an age where sustainability 
is essential, it's a real vision to join forces with initiatives like 
frame.work.

However, for the time being, this vision remains largely inaccessible, 
and is not in line with today's most common needs in the current VMS 
ecosystem.

On the other hand there are real needs for bare metal support in our 
environments. There are even cases where the use of a hypervisor is 
completely impossible. In certain environments, where every IT operation 
must be perfectly calibrated and safe, the use of hypervisors is more 
than ill-received.

It's extremely important to maintain VMS's status as an exception in the 
extreme guarantee of safety, maintainability and longevity. It's 
important that VMS can be used, if need be, as an OS that provides all 
the environment's maintenance without an additional hypervisor layer.

These current practical reasons, combined with a very, very long-term 
vision of proposing an alternative for the laptop, mean that we are very 
sensitive to your arguments.

The unavoidable question is that of platform support and drivers. For 
our current needs in the server environment, our arguments are similar 
to yours, albeit less broadly.

Clearly, our demand for bare metal support will be limited in two ways: 
quantity and time.

For what will be fairly exceptional requests for VMS use in bare metal, 
users won’t be demanding a broad range of choices of options.  One or 
two basic configurations from a reliable manufacturer may suffice.

What's more, given the nature of the demand, it won't be necessary to 
follow the race for new products and device qualification. We're talking 
about applications such as industrial process control and clusters, 
measuring their lifespan in decades. The bare metal chosen in 2024 has 
every chance of being used until 2034.

It would seem that, in a similar way to your analysis, the need for bare 
metal does not require VSI to follow an unsustainable race ahead. And 
the image benefit for the ecosystem is certainly worth the investment.

Again, we agree with your view of the players who, in the long or very 
long term, can bear the investment cost of development and support for 
drivers. For the moment, VMS is a niche market, and bare metal with VMS 
a niche market in a niche market. But those who have the need can 
imagine investment and partnership solutions that would be profitable 
for all. For our needs, there are probably ways to set up collaborations.

Your proposal is also interesting for its dynamics. The return of VMS as 
an alternative must be able to handle different forms of deployment. 
Bare metal (which is the case for 99 % of current VMS instances) is 
certainly one of them, and VSI should be able to support it. Will this 
lead to worldwide success à la Digital? The future will tell.

Thanks again for your attention dear Subcommandante

VMSgenerations