[Oberon] Names and numbers of releases.
Jörg
joerg.straube at iaeth.ch
Fri Jul 17 08:40:38 CEST 2020
> 3) Big data centers, whose HW is beyond the desktop industry.
>
The servers in this space have actaully two goals „increase performance“ and „decrease space“. The space goal is the reason why 19 inch servers are „old fashioned„ and OCP gains momentum. (About 1.5 denser). Obviously energy consumption is a topic as well, but performance and space are more important. So this will fall in my category 2.
> 4) IoT which is more about controlling anything that can be controlled, like video surveillance and all kinds of smart things. Not necessarily mobile.
>
The sensors of IoT are mostly battery-driven and fall in my categoty 1. Here, the power consumption can even be reduced further as „normal“ performance is far too much for sensors and you can decrease performance and with it the power.
> 5) Edge computing, which is the next step beyond IoT. Or perhaps it is IoT under a new name.
>
The word „edge“ has nothing to do with the server HW per se (HW the same as your category 3) above) but with the location where your data center is geographically located. For time critical application, you can not use Amazon‘s data center somewhere in the US, but you have to use computing resources nearer to you, namely at the „edge„ of the Internet. No new HW cycle for this category.
> 6) AI, which is probably much the same as the previous two with some hardware additions dubbed "intelligence".
>
AI needs performance as you have to try a lot of combinations and iterations until you get a decent result. So it falls in my category 2. Obviously mobile devices apply a little bit of AI as well, but if you overdo it, it drains your battery faster.
> 7) Quantum computing which is attracting much attention for a reason that I cannot quite comprehend.
>
This is mainly used to discuss security issues in cryptography. The cryptography guys argument as follows:
- quantum computing (not existing today) will allow for HUGE boost in performance.
- Lets assume quantum computing will exist in 5 years.
- Lets further assume your data you encrypt today is still valid in 5 years, you need to encrypt it „quantum-safe“. If you don‘t do it, a bad guy could steal and store the encrypted data today, and use the upcoming quantum computer in 5 years to decode your important info.
br
Jörg
More information about the Oberon
mailing list