[Info-vax] PowerX Roadmap - Extended beyond 2020

[IanD]

On Saturday, September 10, 2016 at 12:35:05 AM UTC+10, Kerry Main wrote:

<snip>

> Hey, I would love to see VSI do OpenVMS on Power9  at some future point.  As a reminder, VSI has always stated that the X86-64 port was part of the journey, not an end point.
> 
> Having stated this, one should never make a port decision based purely on pure technical reasons.
> 

+1

Digital created some great technologies but when the market changed direction they filed to give customers what they actually wanted

> One needs to "skate where the puck is going to be - not where it is right now" (great Gretzky quote).
> 
> Imho, the industry is moving back to what I would call in simplified terms "a return to client-server". ... but with a slightly different model and emphasis i.e. for lack of a better term "SC-SS" (secure-client to secure-server) ... Back to the future IT.
> 
> By this I mean where the "client" is one of many, many different "secure thin clients" - cell phones, laptops, notebooks, game consoles, and as much as I hate to use industry hype - the "IOT" (fridges, pop machines, scanners .. whatever). 
> 

I may not have been around technology as long as some folk here but since I have been involved (and that started in high school when I worked for 3 months to buy my little casio fx-720P where I learned to program in basic, at a time when computers were something you only did when you got to uni and even then it was for the nerds), computing has been in a continual pendulum swing back and forth between centralization and decentralization, but with each swing filling in more options and offerings with each movement - we will eventually get to a point where the whole spectrum is covered and finally you'll be able to mix n match what you need / want, instead of having to squeeze your technology wants into pre-defined sizes 

I was speaking to an engineer (non IT, military / heavy vehicle design) and he was telling me how much BS he thought the whole IoT was because he didn't want to know how his toaster was 'thinking'. I laughed and told him yeah, on an individual level perhaps he was right but that the real benefit will be around the collation of data coming from say toasters in one suburb and then having that data used by super markets for the prediction of bread sales and better just in time ordering / supplying. I must have triggered something because months later I saw him and he'd been investigating the whole IoT for possible applications in some of the delivery logistics businesses he consults for I think IoT will become rather large as a field not so much because individuals want to know how many liters of water their toilet used that day but because the companies out there want to find some way to monitise all that data to push more crap products down people's throats and/or spin them a tale of how much they can improve their lives by selling them this/that package *sigh*

Not sure where OpenVMS is going to fit in the IoT picture, it's not lean enough or it's file system not quick enough to act as a data collector. Maybe as an aggregator?

> This is a great future market for VSI to market OpenVMS and ARM/X86-64. This is a volume compute focus.
> 

I certainly hope so

If it also runs on hardware that us mere mortals can buy and the license fee's don't require me to sell both kidneys (clustering for example) then I'll be in line to buy

Hobbyists is great and I used it now but I'd be happy to buy a home license much like what MS offer where you can install on up to X computers for their office product

> Wrt to servers, I mean where the "secure server" represents the traditional back end big server environment with lights out, very high HA, multi-site DC models, big core servers with high compute, huge TB memory, ultra-low latency, fast local IO and a reduction of the many different HIGH latency LAN network tiers so common today. Network tier consolidation.
> 

Security is going to become probably one of the main focus points going forward imo. Big business are going to wake up and realize the internet at large is a nasty place and getting worse and no government with their draconian laws and dictates can legislate to fix it. The only option going forward will be for big business to take stuff back under their close guard and to have their own experts actively engaged in protecting their own systems around the clock on a continual basis. I know the drivers for the failure of globalisation are different but this too will drive companies to focus their attention in-house

I was just skimming an article yesterday about the need for next level encryption techniques and the need to start think-tanking solutions. The rationale behind this discussion was that quantum computing will begin to come online in the next decade and with it the treat to existing security techniques. Highly interesting stuff 

OpenVMS-Quantum, lol

> This is a great future market for VSI to market OpenVMS and Power9/X86-64.  This is a mission critical compute focus.
> 
> Why Power9? It's not just because the Power architecture has some really strong differentiators over X86-64 (especially with dedicated very high throughput workload accelerators arch), but also because an improved Power9 partnership with IBM would also get them to more easily justify releasing their many software products on VSI OpenVMS and Power9. 
> 
> >From link provide by Ian:
> https://www.hpcwire.com/2016/08/30/ibm-unveils-power9-details/
> 
> Quote "As we’re moving into the post-Moore’s law era, you can’t just turn the crank and make the general-purpose processor faster,” said Starke. “It’s our believe that you’re going to see more and more specialized silicon. That can be in the form of on-chip acceleration, but as you can see from our approach, we tend to believe it’s more flexible and deployable with off-chip acceleration. Obviously it requires extreme bandwidth, low-latency, and tight integration with your main processor complex, but that’s where we see the future of computing going and you see us putting very strong investments in these directions.” 
> 
> Or stated more simply - skate where the puck is going to be.
> 

Intel have been saying for quite some time now that parallel coding techniques are going to be more important than ever going forward. Just cranking up the clock speed is fast becoming a diminishing return

IBM make some great servers, VSI and IBM might do well together and might slow the wretched Oracle onslaught that's been going on in the marketplace the last decade (I am not a fan of Oracle as you might be able to tell)
 
OpenVMS really needs a strong DB. I doubt Oracle will sell RDB and the other open source db's seem to have stalled on OpenVMS so what to do in this area?
 DB engine rankings show NoSQL still moving forward on market share as is MS SQL Server

> I also think, if implemented, an OpenVMS / Power9 platform would have a much better chance of future success as a mainframe alternative than ANY OS on a X86-64 platform - especially if the IBM LP products were available on the platform to assist with selling Power9 servers. One has to understand the mainframe culture - most really (really) hate all UNIX (including AIX) and Linux and view them as "distributed systems" (mainframe talk for "not real production systems"). On the other hand, mainframe types have a much higher level of respect for OpenVMS. 
> 
> Another good link on Power9 (more technical): August 24, 2016
> http://www.nextplatform.com/2016/08/24/big-blue-aims-sky-power9/ 
> 
> Imho, this would be a great "blue oceans" future strategy.
> 
> Regards,
> 
> Kerry Main
> Kerry dot main at starkgaming dot com

Thanks for those links, I'm heading off to look at them now... :-)