By: Ireland (boh.delete@this.outlook.ie), January 31, 2017 12:51 pm
Room: Moderated Discussions
David Kanter (dkanter.delete@this.realworldtech.com) on January 31, 2017 11:32 am wrote:
> Ireland,
>
> That post was a long rambling off-topic meander about drilling and other stuff.
>
> This is a forum for discussing technology. Keep it on topic.
>
> David
David,
The 'on-topic' part of the contribution, is related to the theory, that one of the earliest applications for the ARM based supercomputer, might actually be in these hostile marine environments, and the like. I.e. Where machines have to be able to run on low power in certain situations, perhaps even off the charge from a solar array. Or on temporary emergency power, in order to make a situation safe, in the case of evacuations etc.
One example, a 'for instance'.
I'm thinking here in particular, of the application of the Java language, where James Gosling has been starting to work on in his career after Sun Microsystems. He's working with these ARM processors already, in that environment. Except, as one scales up to larger installations from the type that he is using at the moment, there might be a need for computing capability on a larger scale, in order to handle more and more variables. If one adds to that then, the idea that Java can be ad-ministered from a remote location and all of the rest of it, there is a entry point for ARM based super-computing, in certain niche applications. I.e. Applications where existing technology based on x86 arrays/servers, and existing software technology - has fallen more or less - flat on it's face, and to the tune of many, many billions of dollars.
I.e. It's not so much the 'cheapness' of the processor technology, that should distract us in this case of the 'Isambard' super-computing project. But the huge down-side risk and cost of systems that fail to work, when they need to. The thing with these applications, supercomputers that are attached to very large pieces of hardware, like a drilling rig - is that it's unlikely to be removed from service for twenty years or more as a piece of hardware. And yet, the hardware and the software are very closely coupled together, and need to be. With something as bog-standard as ARM processors for phones, there might be a potential to create a modular architecture, that could get scaled up over time, even as the hardware, which that supercomputer is trying to control remains the same over a long, extended amount of time. As in, the case of the 'oil rig' example. Or any other kind of remote surveillance hardware, that has to survive against the worst conditions.
In summary.
I think it's one aspect of the technology, that deserves to get a mention. It's also about a larger responsibility that software engineers have, as more and more software interacts with a lot more physical infrastructure. We're unlikely to get to the bottom of it today, or tomorrow. It's something we need to think about.
> Ireland,
>
> That post was a long rambling off-topic meander about drilling and other stuff.
>
> This is a forum for discussing technology. Keep it on topic.
>
> David
David,
The 'on-topic' part of the contribution, is related to the theory, that one of the earliest applications for the ARM based supercomputer, might actually be in these hostile marine environments, and the like. I.e. Where machines have to be able to run on low power in certain situations, perhaps even off the charge from a solar array. Or on temporary emergency power, in order to make a situation safe, in the case of evacuations etc.
One example, a 'for instance'.
I'm thinking here in particular, of the application of the Java language, where James Gosling has been starting to work on in his career after Sun Microsystems. He's working with these ARM processors already, in that environment. Except, as one scales up to larger installations from the type that he is using at the moment, there might be a need for computing capability on a larger scale, in order to handle more and more variables. If one adds to that then, the idea that Java can be ad-ministered from a remote location and all of the rest of it, there is a entry point for ARM based super-computing, in certain niche applications. I.e. Applications where existing technology based on x86 arrays/servers, and existing software technology - has fallen more or less - flat on it's face, and to the tune of many, many billions of dollars.
I.e. It's not so much the 'cheapness' of the processor technology, that should distract us in this case of the 'Isambard' super-computing project. But the huge down-side risk and cost of systems that fail to work, when they need to. The thing with these applications, supercomputers that are attached to very large pieces of hardware, like a drilling rig - is that it's unlikely to be removed from service for twenty years or more as a piece of hardware. And yet, the hardware and the software are very closely coupled together, and need to be. With something as bog-standard as ARM processors for phones, there might be a potential to create a modular architecture, that could get scaled up over time, even as the hardware, which that supercomputer is trying to control remains the same over a long, extended amount of time. As in, the case of the 'oil rig' example. Or any other kind of remote surveillance hardware, that has to survive against the worst conditions.
In summary.
I think it's one aspect of the technology, that deserves to get a mention. It's also about a larger responsibility that software engineers have, as more and more software interacts with a lot more physical infrastructure. We're unlikely to get to the bottom of it today, or tomorrow. It's something we need to think about.