Experimental Physics and
Industrial Control System

Matthias Clausen <[email protected]> · Tue, 22 Feb 2005 09:00:20 +0100

Hi Klemen,
fault tolerance and reliability are topics which will be also discussed
for the next EPICS version-V4.

Within the time frame of the next three years we are planning to
introduce redundant IOC's into the next EPICS version. In order to
introduce the collected ideas of the team working on these subjects -
rather than individual thoughts - it will be necessary to know your time
frame for the survey.

I'll be happy to enter our current thoughts but I would prefer to enter
them based on an on going feasibility study and the results of several
EPICS meetings within the next months.
Please keep me informed. It seems we have a common ground here.
-Matthias
Klemen Zagar wrote:
Dear EPICS users,
Cosylab is working on a EU-funded project on "Dependable Distributed
Systems" (DeDiSys -- http://www.dedisys.org). The goal of the project
is to investigate and improve availability of distributed systems when
faults are present (e.g., when the network is down or when nodes crash).
Since EPICS is an infrastructure for distributed systems, we are
considering to apply the results of our research work to EPICS as
well. In order to do that, we need input from those who use EPICS in
production environments, such as yourselves.
We have prepared a general questionnaire that is expected to gauge
requirements and expectations with regard to fault tolerant behavior.
I kindly ask you too take a bit of your time and fill it in. It is
conveniently available on the web:
 http://www.xlab.si/dedisys/start.php
(It takes about 20 to 30 minutes to make it through all the questions.
If you wish to be kept informed of the project's progress, you might
also wish to sign-in for the interest group.)
Since questionnaire is very generic, we would also welcome any
additional input specific to EPICS, such as:
1. How frequent are failures at your facility?
2. How many of the failures are critical in a sense that operations of
the entire facility are disturbed (e.g., a beam dump)?
3. How many failures could be attributed to EPICS or the control
system itself (e.g., network congestions, defects in software, erratic
behavior of controls hardware, ...)?
4. How many of these failures could EPICS help to avoid, and in what
way (redundancy, replication of records, reliability APIs for
developers, guidelines for development of reliable systems, ...)?
5. What is the typical down-time as a sequence of a failure?
6. Could EPICS be improved to help reduce this down-time? If so, how
(e.g., allow reconfiguration without IOC restart, ...)?
In order to reduce traffic on the mailing list, please answer directly
to myself, and I will summarize your contributions in a later posting.
Thank you for your cooperation,
  Klemen Zagar.

--
------------------------------------------------------------------------
Matthias Clausen                         Cryogenic Controls Group(MKS-2)
phone:  +49-40-8998-3256                Deutsches Elektronen Synchrotron
fax:    +49-40-8994-3256                                    Notkestr. 85
e-mail: [email protected]                           22607 Hamburg
WWW-MKS2.desy.de                                                 Germany
------------------------------------------------------------------------

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System