Re: newlib hours in service

[Joel Sherrill <joel dot sherrill at oarcorp dot com>]

On 9/6/2017 10:24 AM, Alan Ludwig wrote:
> Hello,
>
> I'm trying to get an estimate of the order of magnitude of the total number of hours in service for newlib.  Is it thousands of hours? Millions of hours? Tens of millions of hours?  For which versions?

Given that newlib is primarily used with Cygwin, RTEMS, and bare
metal toolsets, I would assume millions of hours.

RTEMS is an free embedded RTOS that heavily used in the space
and scientific applications. It has many commercial users but
space and science publish papers. We have been using newlib
since the early 1990s. Every RTEMS application is a newlib
application. Translating that into hours of run-time is hard
but here are some sample currently active system

+ Electra radio on Curiosity Rover (5+ years)
+ Fermi Gamma Ray Burst Detector (9+ years since launch)
+ Solar Dynamics Observatory (7+ years)
+ Magnetospheric Multiscale Mission (2+ years w/ 4 satellites)
+ Galileo GPS Constellation (all CPUs on all satellites)
+ ESA Herschel (decommissioned ~4 years)
+ ESA Planck (decommissioned ~4 years)
+ NASA Juno (6+ years to Jupiter, probably slept on way)
+ Stanford Linear Accelerator (? years and number of nodes)

That ignores commercial systems which I can't speak of or
don't know about. I know of one specialized vehicle system
which had about 1200 fielded by 2002 with an unknown average
number of "operational hours" per unit. But they had to be
used. Even 10 hours a week per unit for 15 years is 180,000
hours for that system alone.

RTEMS is also deployed by one user in building control systems.
That's a 24/7/365 system and I have no idea how many they have
sold. But they are well known.

From RTEMS alone, I would pretty confidently say there are
millions of hours. But each user would only be using a
differing subset of newlib.

Add in Cygwin and other newlib users, and there have to be
millions of hours.

> Usually, I'd work this backwards from estimates of the number of downloads, installed seats, size of the user base, etc. However, newlib doesn't seem to have any of this sort of information available at the download or support websites.

For the past few years, the RTEMS Community has been encouraged
to build from source and fetch it from the appropriate place
using a Python based builder. Before that, we built binaries
and distributed those. So the RTEMS Community contribution to
downloads for a long time won't be present. But someone from
sourceware.org may be able to help for what they can account for.
 
> Do you have any information that I might use to estimate the hours in service for newlib?  The numbers can be very rough. I'm only trying to choose an order of magnitude.  Is it 10^4 or say 10^7 total hours?  Any information that I might use to make these gross estimates would be appreciated.
>
> Thanks,
>
> Alan

--
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35806
Support Available                (256) 722-9985