[Starlink] [E-impact] DataCenters in Space (was Re: fiber IXPs in space)

Ulrich Speidel u.speidel at auckland.ac.nz
Thu Apr 20 17:21:21 EDT 2023


Indeed. There's another point that's been missed in the "superconductor" 
suggestion: Why do we get the heat in the first place?

Superconductors are great when it comes to reducing resistive losses in 
long and / or high current conductors (power distribution, MRI magnets, 
...). But this isn't why computer chips get hot. Let me take you back to 
your Physics 101 when you learned that Power P was the product of 
current I and voltage V. A logic chip like a CPU is nothing but an 
assortment of gazillions of little switches. When a switch is open, it 
may have voltage across it but no current flows: no power gets 
dissipated. If it's closed, current may flow but there won't be any 
voltage across it. Also not a source of power loss.

The power loss (heat generation) happens when the little switches are 
switching, i.e., when they are between open and closed and when there is 
both a bit of voltage and a bit of current present. Naively you might 
say that a switch is either on or off, and so that shouldn't occur, but 
in both theory and practice, an instantaneous loss-free switching 
process requires a signal of infinite bandwidth when subjected to 
Fourier analysis. Fourier analysis allows us to model any signal as a 
combination of sinusoidal signals of different frequency, amplitude and 
phase, and it's in particular the high-and-in-the-direction-of-infinity 
frequency components of that combination that are needed for the 
"ïnstantaneous"switching. Unfortunately, in any real circuit of larger 
than zero size, reactive elements (capacitive and inductive components 
or parasitic properties of that nature) attenuate these. So the only 
real switching we can actually do in real life is switching that 
dissipates power when it happens - no matter whether the chip is built 
using superconductors or not.

In a modern CPU, a significant percentage of gates are this this "gray" 
in-between state between 0 and 1 for a significant part of the time, 
which is why you need elaborate cooling fans and water coolers etc., and 
it's also why clock frequencies haven't increased substantially in 
recent years.

On 21/04/2023 12:06 am, Dave Collier-Brown via Starlink wrote:
>
> Another point they missed: on earth, we can use conductive cooling and 
> transfer the heat from the machines to a flow of air.  In space, we 
> can only use radiative cooling, and we need to be out of the sun to 
> have enough temperature difference.
>
> --dave
>
> On 4/20/23 07:10, Hesham ElBakoury via Starlink wrote:
>> The article about the ASCEND project says:
>> "Very low ambient temperatures in space will dramatically reduce the 
>> need for cooling equipment that consumes enormous amounts of energy. 
>> A significant part of a data center’s energy use is for cooling 
>> equipment, accounting for more than 50% in some facilities. 
>> Temperatures can be as low as -292°F (-180°C) when an orbiting object 
>> is in the Earth’s shadow."
>>
>> Hesham
>>
>> On Wed, Apr 19, 2023, 10:44 PM Daniel Schien 
>> <Daniel.Schien at bristol.ac.uk> wrote:
>>
>>     I assume any object in orbit will be hidden from the sun some of
>>     the time. So, the machines will require some pretty big battery
>>     to go up with them.
>>
>>     I'd like to also know what the launch cost is.
>>
>>     Tom Segert estimates in his LinkedIn post, for a 100kg satellite
>>     payload:
>>
>>     "TL:DR ~57 ton CO2e for a typical ESA satellite (including Ariane
>>     6 launch), <15t CO2e for a satellite built in a factory and
>>     launched with a re-usable rocket."
>>
>>     Depending on the type of server that should go up there, this is
>>     a fair amount of carbon to offset from brighter sunlight.
>>
>>     The article also gets the carbon footprint wrong:
>>
>>     "Data centers are big energy consumers – between 2% and 3% of all
>>     global consumption – a rate that is doubling every year."
>>
>>     The latest was IEA estimating it to be around 220-320 TWh (out of
>>     30,000) in 2021 data and growing between 10-60% over 6 years in
>>     total (so let's than 10 CAGR). But it's certainly not doubling
>>     every year. That's just completely wrong.
>>
>>
>>           DanielSchien
>>
>>     Senior Lecturer in Computer Science
>>
>>     Department of Computer Science | University of Bristol
>>
>>     	
>>     	
>>     	
>>
>>     *Submit software engineering project ideas for 2022*
>>
>>
>>     	http://bris.ac.uk/software-engineering
>>     Watch: https://youtu.be/lU-ZsBDFWDI
>>     <https://youtu.be/lU-ZsBDFWDI>
>>
>>
>>     	Merchant Venturers Building , Woodland Rd Bristol, BS8 1UB
>>
>>     *Book a meeting*:
>>     https://outlook.office365.com/owa/calendar/OfficeHours@bristol.ac.uk/booki
>>     <https://outlook.office365.com/owa/calendar/OfficeHours@bristol.ac.uk/bookings/>
>>
>>
>>
>>     ------------------------------------------------------------------------
>>     *From:* E-impact <e-impact-bounces at ietf.org> on behalf of Vint
>>     Cerf <vint=40google.com at dmarc.ietf.org>
>>     *Sent:* Thursday, April 20, 2023 2:16:38 AM
>>     *To:* tom at evslin.com <tom at evslin.com>
>>     *Cc:* Michael Richardson <mcr at sandelman.ca>; starlink
>>     <starlink at lists.bufferbloat.net>; e-impact at ietf.org
>>     <e-impact at ietf.org>
>>     *Subject:* Re: [E-impact] [Starlink] DataCenters in Space (was
>>     Re: fiber IXPs in space)
>>     O&M will be a bear
>>     v
>>
>>
>>     On Wed, Apr 19, 2023 at 9:13 PM Tom Evslin via Starlink
>>     <starlink at lists.bufferbloat.net> wrote:
>>
>>         I think space-based data centers will be the rule rather than
>>         the exception. Wrote about that a couple of years ago
>>         although, as usual, things have not happened as quickly as I
>>         predicted
>>         https://blog.tomevslin.com/2021/07/computing-clouds-in-orbit-a-possible-roadmap.html
>>
>>         -----Original Message-----
>>         From: Starlink <starlink-bounces at lists.bufferbloat.net> On
>>         Behalf Of Michael Richardson via Starlink
>>         Sent: Wednesday, April 19, 2023 7:35 PM
>>         To: starlink <starlink at lists.bufferbloat.net>; e-impact at ietf.org
>>         Subject: [Starlink] DataCenters in Space (was Re: fiber IXPs
>>         in space)
>>
>>
>>         I saw this reported in BIS-Spaceflight.
>>         (I'm usually a few months behind in reading it) I like the
>>         "first objective"!
>>
>>         https://www.thalesgroup.com/en/worldwide/space/press-release/ascend-thales-alenia-space-lead-european-feasibility-study-data
>>
>>         Cannes, November 14, 2022 – Thales Alenia Space, the joint
>>         company between Thales (67%) and Leonardo (33%), has been
>>         chosen by the European Commission to lead the ASCEND
>>         (Advanced Space Cloud for European Net zero emission and Data
>>         sovereignty) feasibility study for data centers in orbit, as
>>         part of Europe’s vast Horizon Europe research program.
>>
>>         Digital technology’s expanding environmental footprint is
>>         becoming a major
>>         challenge: the burgeoning need for digitalization means that
>>         data centers in Europe and around the world are growing at an
>>         exponential pace, which in turn has a critical energy and
>>         environmental impact.
>>
>>         The first objective of this study will be to assess if the
>>         carbon emissions from the production and launch of these
>>         space infrastructures will be significantly lower than the
>>         emissions generated by ground-based data centers, therefore
>>         contributing to the achievement of global carbon neutrality.
>>         The second objective will be to prove that it is possible to
>>         develop the required launch solution and to ensure the
>>         deployment and operability of these spaceborne data centers
>>         using robotic assistance technologies currently being
>>         developed in Europe, such as the EROSS IOD demonstrator.
>>
>>         This project is expected to demonstrate to which extent
>>         space-based data centers would limit the energy and
>>         environmental impact of their ground counterparts, thus
>>         allowing major investments within the scope of Europe’s Green
>>         Deal, possibly justifying the development of a more
>>         climate-friendly, reusable heavy launch vehicle. Europe could
>>         thus regain its leadership in space transport and space
>>         logistics, as well as the assembly and operations of large
>>         infrastructures in orbit.
>>
>>         _______________________________________________
>>         Starlink mailing list
>>         Starlink at lists.bufferbloat.net
>>         https://lists.bufferbloat.net/listinfo/starlink
>>
>>         _______________________________________________
>>         Starlink mailing list
>>         Starlink at lists.bufferbloat.net
>>         https://lists.bufferbloat.net/listinfo/starlink
>>
>>
>>
>>     -- 
>>     Please send any postal/overnight deliveries to:
>>     Vint Cerf
>>     Google, LLC
>>     1900 Reston Metro Plaza, 16th Floor
>>     Reston, VA 20190
>>     +1 (571) 213 1346
>>
>>
>>     until further notice
>>
>>
>>
>>     -- 
>>     E-impact mailing list
>>     E-impact at ietf.org
>>     https://www.ietf.org/mailman/listinfo/e-impact
>>
>>
>> _______________________________________________
>> Starlink mailing list
>> Starlink at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/starlink
> -- 
> David Collier-Brown,         | Always do right. This will gratify
> System Programmer and Author | some people and astonish the rest
> dave.collier-brown at indexexchange.com  |              -- Mark Twain
>
> */CONFIDENTIALITY NOTICE AND DISCLAIMER/*/ : This telecommunication, 
> including any and all attachments, contains confidential information 
> intended only for the person(s) to whom it is addressed. Any 
> dissemination, distribution, copying or disclosure is strictly 
> prohibited and is not a waiver of confidentiality. If you have 
> received this telecommunication in error, please notify the sender 
> immediately by return electronic mail and delete the message from your 
> inbox and deleted items folders. This telecommunication does not 
> constitute an express or implied agreement to conduct transactions by 
> electronic means, nor does it constitute a contract offer, a contract 
> amendment or an acceptance of a contract offer. Contract terms 
> contained in this telecommunication are subject to legal review and 
> the completion of formal documentation and are not binding until same 
> is confirmed in writing and has been signed by an authorized signatory./
>
>
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
-- 
****************************************************************
Dr. Ulrich Speidel

School of Computer Science

Room 303S.594 (City Campus)

The University of Auckland
u.speidel at auckland.ac.nz  
http://www.cs.auckland.ac.nz/~ulrich/
****************************************************************


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/starlink/attachments/20230421/e0a5d277/attachment-0001.html>


More information about the Starlink mailing list