walden@dip.eecs.umich.edu (Eugene Marvin Walden) (08/10/90)
Well, folks, this is part II of my summary of commercial real-time operating
systems. I was able to find most of the addresses, but, as you can see, six are
still missing. Anyway, if anybody has corrections or further additions, please
E-mail me.
I have tried to provide a rough classification of the RTOSs that I have
seen. The classifications may be inaccurate, or just plain wrong, but I tried.
After I receive literature from these companies, maybe I'll post a summary of
each OS. Have fun!
1. Regulus
Alcyon Corp.
6888 Nancy Ridge Dr.
San Diego, CA. 92121
(619) 587-1155
- UNIX-like operating system with real-time extensions
2. REAL/IX
Modcomp
John Worden
1650 W. McNab Rd.
Ft. Lauderdale, FL 33340-6099
(305) 977-1083
- UNIX-like operating system with real-time extensions
3. CX/RT
Harris Computer Systems Division
2101 W. Cypress Creek Rd.
Ft. Lauderdale, FL.
(305) 974-1700
- UNIX-like operating system with real-time extensions
- Scheduler is a combination of a deadline-driven scheduler and a multi-
level round-robin scheduler
4. Chimera II
Carnegie-Mellon University
- UNIX-like operating system with real-time extensions
5. LynxOS
Lynx Real-Time Systems, Inc.
550 Division St.
Campbell, CA. 95008
(408) 370-2233
- UNIX-like operating system with real-time extensions
6. FlexOS
Digital Research, Inc.
Integrated Systems Business Unit
70 Garden Court, B206
Monterey, CA 93940
(408) 982-0700
- UNIX-like operating system with real-time extensions
7. RTE
Hewlett-Packard
8. pSOS
Software Components Group
1731 Technology Dr.
San Jose, CA 95110
(408) 437-0700
- UNIX-compatible kernel, < 12K
9. MTOS (Industrial Programming, Inc.)
10. PDOS (Eyering Research)
11. C Executive
JMI Software Consultants
904 Sheble Ln. P.O. Box 481
Spring House, PA 19477
(215) 628-0846
12. RTUX
Emerge Systems
Frank Aaron
114 6th Ave., P.O. Box 3175
Indialantic, FL 32903
(407) 723-0444
- UNIX-like operating system with real-time extensions
13. CSX (Cytek)
14. AMX (Kadak)
Kadak Products, Ltd.
206-1847 West Broadway
Vancouver, BC Canada V6J 1Y5
(604) 734-2796
15. UniFlex
UniFLEX Computing, Ltd.
111 Providence Rd.
Chapel Hill, NC 27514
(919) 493-1451
Sales: (800) 486-1000
16. IDRIS (Whitesmiths)
- UNIX-like operating system with real-time extensions
17. RTXC
A. T. Barrett & Assoc.
Ron Hodge
11501 Chimney Rock, Suite R
Houston, TX 77035
18. USX
US Software
John Bayer
14215 NW Science Park Dr.
Portland, OR 97229
(503) 641-8446
19. UMAX V OS
Encore Computer Corp.
6901 W. Sunrise Blvd.
Ft. Lauderdale, FL 33313
(305) 587-2900
- UNIX-like operating system with real-time extensions
20. VRTX
Ready Systems
470 Potrero Ave.
P.O. Box 60217
Sunnyvale, CA 94088-0217
(800) 228-1249
- small, ROMable kernel
- priority-based round-robin scheduler
- message passing, interrupt handling, semaphores
21. VxWorks
Wind River Systems
1351 Ocean Ave.
Emeryville, CA 94608
(415) 428-2623
- small, ROMable kernel
- priority-based round-robin scheduler
- message passing, interrupt handling, semaphores
22. iRMX and Distributed iRMX
Intel Corp.
5200 N.E. Elam Young Pkwy.
M.S. HF3-58
Hillsboro, OR 97124
(503) 681-8080
- small, ROMable kernel
- priority-based round-robin scheduler
- message passing, interrupt handling, semaphores
23. QNX
Quantum Software Systems, Ltd.
175 Terrence Matthews Crescent
Kanata, Ontario Canada K2M 1W8
(613) 591-0931
- like UNIX, but different enough to annoy the hell out of you
- priority-based round-robin scheduler
- message passing, interrupt handling, semaphore locking
24. Alpha
Concurrent Computer Corp.
One Technology Way
Westford, MA 01886
(508) 692-6200
- two-dimensional priority specification
- real-time transactions
- object-oriented structure
-
25. OS9, OS9000
Microware
1900 N.W. 114th St.
Des Moines, Iowa 50322
(515) 224-1929
26. RTU
Concurrent Computer Corp.
One Technology Way
Westford, MA 01886
(508) 692-6200
27. RT-11
Digital Equipment Corp.
Class 1:
The oldest "real-time" operating systems are systems that were used back in
the Dark Ages such as Hewlett-Packerd's RTE and DEC's RT-11. Keep in mind that
I never had to use the dreaded punch cards, so Dark Ages to me means early
eighties :-).
Class 2:
The second generation of "real-time" operating systems were the real-time
executives that were used for embedded data acquisition & control. These exec-
utives typically handle message-passing primitives, simple task scheduling
according to priority, semaphores and queues. They are usually fairly small,
the main purpose being to provide an interface to a single-board computer. Ex-
amples of such real-time operating systems are:
- VRTX
- iRMX
- VxWorks
- AMX
- C-Executive
Class 3:
With the rebirth of UNIX in the eighties, there also came an entourage of
companies trying to retro-fit UNIX for real-time applications. Usually, the
modifications include:
- Fully preemptive, priority-based schedulers
- Ability to lock tasks in memory (i.e. no page-faults for critical tasks)
- Faster context switches
- Lower interrupt latency
- Higher granularity
Examples include:
- Regulus
- REAL/IX
- LynxOS
- FlexOS
- pSOS
- RTUX
- QNX
- Idris
Class 4:
This class is closely related to Class 3 above. More companies are introduc-
ing multiprocessor machines, and with these machines come new operating sys-
tems. All such systems I have seen are some form of UNIX, but extended to
handle multiprocessing and real-time. Examples include:
- CX/RT
- UMAX V OS
- RTU
Class 5:
I am biased towards the RTOSs with the academically interesting schedulers,
but currently, there are not many. Some of the UNIX-like RTOSs are working on
schedulers other than priority-based schedulers.
1. Alpha has a modular scheduler. Policy modules can be inserted into the
scheduling scheme, so Alpha itself does not dictate the scheduling pol-
icy. I believe that the available scheduling policies are simple priority
based schedulers, deadline-driven schedulers, and possibly the Best
Effort Scheduler.
2. LynxOS is developing a deadline-driven scheduler for the NASA Space
Station.
3. CX/RT uses some combination of a deadline-driven scheduler and a multi-
level round robin scheduler.
4. Chimera-II uses a least-laxity scheduler
There you have it. If anybody wants to send me info, my mailing address is:
Eugene Walden
University of Michigan
Space Physics Research Lab
2455 Hayward St.
Ann Arbor, MI. 48109
(313) 764-0137
walden@dip.eecs.umich.edusrp@modcomp.UUCP (Steve Pietrowicz) (08/12/90)
in article <3036@zipeecs.umich.edu>, walden@dip.eecs.umich.edu (Eugene Marvin Walden) says:
] REAL/IX
]
] Modcomp
] John Worden
] 1650 W. McNab Rd.
] Ft. Lauderdale, FL 33340-6099
] (305) 977-1083
]
] - UNIX-like operating system with real-time extensions
Actually "UNIX-like" is incorrect. REAL/IX is *real* AT&T UNIX, with real-time
extensions. It has not been re-written from scratch as some places claim you
have to do to get real-time. Being real AT&T is also one of the things (besides
it's real-time features) that makes REAL/IX attractive to companies that
license it.
Thanks for the mention in your list!
Stevejlaurent@encore.com (Jim Laurent) (08/13/90)
> 19. UMAX V OS > > Encore Computer Corp. > 6901 W. Sunrise Blvd. > Ft. Lauderdale, FL 33313 > (305) 587-2900 > > - UNIX-like operating system with real-time extensions > I'd like to make a correction. UMAX V is Encore's port of ATT System V OS not with real time extensions but with extensions to host a co-resident RT OS called Micro-MPX. Micro-MPX is a true real time kernal with very fast response times due to its implementation of light-weight threads of execution. Micro-MPX is a redesign of MPX-32, Encore's proprietary real-time OS for its CONCEPT-32 computer systems. Micro-MPX support MPX-32 compatibility along with a POSIX 1003.4 interface. Co-resident operating systems allow the creation of multiple virtual machines or domains within a multiprocessor. Both UMAX and Micro-MPX are fully multithreaded in support of the 90 family's symmetrical multiprocessor design. Micro-MPX will be available on Encore's recently announced 91 Series of 88k base computers. Jim Laurent UUCP: uunet!gould!jlaurent Greenbelt, MD Office INTERNET: jlaurent@encore.com Encore Computer Corporation
unhd (Roger Gonzalez ) (08/13/90)
>8. pSOS > > Software Components Group > 1731 Technology Dr. > San Jose, CA 95110 > (408) 437-0700 > > - UNIX-compatible kernel, < 12K ^^^^^^^^^^^^^^^^^^^^^^ No it isn't. Not even remotely. It's small, fast, but crippled. We've spent more time trying to *make* it Unix compatible than I care to think about. We're moving to VxWorks, with luck. No SCG rebuttal mail, thank you. -Roger -- UUCP: ..!uunet!unhd!rg | USPS: Marine Systems Engineering Laboratory BITNET: r_gonzalez at unhh | University of New Hampshire PHONE: (603) 862-4600 | Marine Programs Building FAX: (603) 862-4399 | Durham, NH 03824-3525
bill@mwca.UUCP (Bill Sheppard) (08/13/90)
In article <3036@zipeecs.umich.edu=> walden@dip.eecs.umich.edu (Eugene Marvin Walden) writes:
=>
=> Well, folks, this is part II of my summary of commercial real-time operating
=>systems...
=>
=>25. OS-9, OS-9000
=>
=> Microware
=> 1900 N.W. 114th St.
=> Des Moines, Iowa 50322
=> (515) 224-1929
=>
- small, ROMable kernel
- pre-emptive, priority-based round-robin scheduler
- signals, events (semaphores), pipes/named pipes, data modules
- full resident or cross development
=>...
=> The second generation of "real-time" operating systems were the real-time
=>executives that were used for embedded data acquisition & control. These exec-
=>utives typically handle message-passing primitives, simple task scheduling
=>according to priority, semaphores and queues. They are usually fairly small,
=>the main purpose being to provide an interface to a single-board computer. Ex-
=>amples of such real-time operating systems are:
=>...
OS-9 is one of the earliest RT OS's (at least for microprocessor-based
systems) having been first released in 1979 for the 6809. Because of
its original design as a full-blown operating system rather than a
stripped-down kernal, however, it tends to be quite similar to UNIX,
allowing full resident development in a very UNIX-like environment
rather than requiring cross-development. Its modular design does allow
for ROM-based systems in the same size and offering similar performance
to the RT kernels.
=>...
=> Eugene Walden
=> walden@dip.eecs.umich.edu
--
################################################################################
# Bill Sheppard -- bill@mwca.microware.com -- {uunet,sun}!mcrware!mwca!bill #
# Microware Systems Corporation --- OS-9: Seven generations beyond __/_!! #
#######Opinions expressed are my own, though you'd be wise to adopt them!#######rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) (08/16/90)
In article <1990Aug13.123602.509@uunet!unhd> rg@unhd.UUCP (Roger Gonzalez ) writes: > >>8. pSOS >> >> - UNIX-compatible kernel, < 12K > ^^^^^^^^^^^^^^^^^^^^^^ >No it isn't. Not even remotely. It's small, fast, but crippled. We've >spent more time trying to *make* it Unix compatible than I care to think >about. We're moving to VxWorks, with luck. > >No SCG rebuttal mail, thank you. >-Roger Well I don't work for Software Components Group so I feel I can rebut. pSOS is not intended to be Unix compatible and if Roger wasted time trying to make it so then he is the one who made the mistake. pSOS absolutely is not 'crippled' and is an excellent real time kernel (from experience) and I have seen at least one independent, indepth review of real time kernels that ranked it as the best. -- Jim Nusbaum, Computer Research Lab, Tektronix, Inc. [ucbvax,decvax,allegra,uw-beaver,hplabs]!tektronix!tekchips!rjn rjn@tekchips.labs.tek.com (503) 627-4612
abg@stc06.ornl.gov (BANGS A L) (08/17/90)
In article <6651@tekcrl.LABS.TEK.COM> rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) writes: >In article <1990Aug13.123602.509@uunet!unhd> rg@unhd.UUCP (Roger Gonzalez ) writes: >> >>>8. pSOS >>> >>> - UNIX-compatible kernel, < 12K >> ^^^^^^^^^^^^^^^^^^^^^^ >>No it isn't. Not even remotely. It's small, fast, but crippled. We've >>spent more time trying to *make* it Unix compatible than I care to think >>about. We're moving to VxWorks, with luck. >> Ok, time for my two-bits. I have been looking very closely at VMEexec, which gives you the power of the pSOS+ kernel with some nice stuff added by Motorola, including SVIDlib, a Unix library, and communications to make ethernet and backplace communications transparent. It seems that SVIDlib would make pSOS a tolerable product for those of us who are looking to do Unix development but want to run the final product in real-time. Alex Alex L. Bangs ---> bangsal@ornl.gov Of course, my opinions are Oak Ridge National Laboratory/CESAR my own darned business...
ghfeil@white.toronto.edu (Georg Feil) (08/21/90)
In article <6651@tekcrl.LABS.TEK.COM> rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) writes: >In article <1990Aug13.123602.509@uunet!unhd> rg@unhd.UUCP (Roger Gonzalez ) writes: >>>8. pSOS >>... >>No it isn't. Not even remotely. It's small, fast, but crippled. We've >>... > >pSOS absolutely is not 'crippled' and is an excellent real time kernel >(from experience) and I have seen at least one independent, indepth >review of real time kernels that ranked it as the best. >... > You're both right. The issue here, I think, is the difference between a kernel and its development environment, which may include run-time libraries which are Unix-compatible. Georg. -- Georg Feil Internet: ghfeil@white.toronto.edu -or- : georg@sgl.ists.ca ..if all else fails, try: {uunet,pyramid,watmath,utzoo}!utcsri!white!ghfeil (UUCP) ghfeil%white.toronto.edu@relay.cs.net (ARPA)
rjsnyder@cbnewsk.att.com (robert.snyder) (08/22/90)
In article <6651@tekcrl.LABS.TEK.COM>, rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) writes: > In article <1990Aug13.123602.509@uunet!unhd> rg@unhd.UUCP (Roger Gonzalez ) writes: > > > >>8. pSOS > >> [...lines deleted...] > ...and I have seen at least one independent, indepth > review of real time kernels that ranked it as the best. Where and when was the review? R.J. Snyder
rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) (08/22/90)
In article <1990Aug21.175052.18395@cbnewsk.att.com> rjsnyder@cbnewsk.att.com (robert.snyder) writes: >In article <6651@tekcrl.LABS.TEK.COM>, rjn@snowbird.LABS.TEK.COM (Jim Nusbaum) writes: >> ...and I have seen at least one independent, indepth >> review of real time kernels that ranked it as the best. > >Where and when was the review? > > R.J. Snyder Since I have had a couple of requests for this: the review was an internal document done by the control systems group at Fermi National Accelerator Lab in Batavia, IL. I believe it was done about three years ago now (that was when I was there and I think it was new when I saw it) and compared pSOS, VRTX, OS9 and others (I think). By now it is probably outdated. I do not have a copy and can not give any more detailed pointers on how to get a copy. -- Jim Nusbaum, Computer Research Lab, Tektronix, Inc. [ucbvax,decvax,allegra,uw-beaver,hplabs]!tektronix!tekchips!rjn rjn@tekchips.labs.tek.com (503) 627-4612