[evla-sw-discuss] Next MIB mtng, time & place

[Bill Sahr]

I have reserved the Upper Conference Room for Wednesday,
2/6/2002 at 1:30PM for the next meeting of the full group
to discuss MIB chips, boards, and RTOSes.  We have the
room until 3:30PM if we need it.  Please let me know if 
this date &/or time is problematic for any of you.

Thanks.

Bill


-------- Original Message --------
Subject: Next MIB meeting
Date: Mon, 21 Jan 2002 18:44:30 -0700
From: Bill Sahr <bsahr at cv3.cv.nrao.edu>
Organization: National Radio Astronomy Observatory
To: Bruce Rowen <browen at cv3.cv.nrao.edu>, Kevin Ryan
<kryan at cv3.cv.nrao.edu>,Bill Sahr <bsahr at cv3.cv.nrao.edu>,Ken Sowinski
<ksowinski at cv3.cv.nrao.edu>,Barry Clark <bclark at cv3.cv.nrao.edu>,Wayne
Koski <wkoski at cv3.cv.nrao.edu>,George Peck <gpeck at cv3.cv.nrao.edu>,Brent
Carlson <Brent.Carlson at nrc.ca>,Jim Jackson <jjackson at cv3.cv.nrao.edu>

I am trying for a room reservation for Monday, 2/4/2002 or
Tuesday, 2/5/2002 in the upper conference room as the date
and place for the next meeting to discuss chip & RTOS
candidates.  The goal is to have between 1 - 5 "chips" and
a few RTOSes.

Chip Characteristics
- as much on-chip RAM as possible, preferably => 512 kbytes
- low or variable clock speed
- SPI or a serial port configurable as such
- watchdog timer
- at least 1 additional timer
- parallel I/O
- interface to flash RAM, to hold diagnostics

In the selection of the chip, RFI is a dominant consideration
and the major reason that we are looking for as much on-chip
RAM as possible.  We are trying to avoid executing programs
which reside in off-chip RAM, except for the case of diagnotics.


RTOSes
We are trending toward something with a small footprint,
highly modular, with a TCP/IP stack.

Possibilities
- ATI Nucleus
- SuperTask!  (US Software, a Lantronix company)
  We need cost & configuration data, and a technical
  evaluation of this one.
The above two are under consideration since they were the 
"recommended" RTOSes for the EC-1 (now DSTni) chip

- Is VxWorks useable in the emerging context ?
  What is its minimum footprint, with the TCP/IP stack ?
- QNX, same questions as for VxWorks
- pSOS, same questions as for VxWorks

Any other good, solid candidates ?  What about RTEMS ?
What about microCLinux ?

I think we are all agreed that those varieties of real-time
Linux, such as RTLinux and RedSonic's REDICE-Linux 2.0,
which offer a fullup Linux are probably inappropriate
for an embedded system where on-chip RAM is likely to 
be at a scarce resource.