Re: History of EPICS

[bob dalesio via Tech-talk <tech-talk at aps.anl.gov>]

Hi Dimitry,

As you ask and I am an old guy that enjoys stories...

I am attaching an early paper on the process database structure. 

Is there any paper describing "Why is EPICS the way it is?"?  I.e., what
were the reasons to design EPICS around the "database concept" (and was
it designed like this from the very beginning or if that was done later)?

GTACS had two major influences early on - network communication design that was done by Jeff Hill with work started on the Accelerator Control System - that became the Vsystem product, and emerging industrial control system tools which came out of Foxboro and featured a process database. I worked at two companies that were derived from Foxboro - EMC Controls with the D3 product (that ran the DESY cryo system) and Computer Products Inc that sold fully redundant hardware for nuclear power plants. 

The innovation came with standardizing the client on a narrow interface that defined a Process Variable as a value of some data type along with the time stamp and alarm status of that data (DBR_Types) and making all hardware communication a layer below the process database - so that all record processing did not wait for IO to complete - which made records process at the frequency specified - which is crucial to performing PID control. Marty Kraimer broke the interface to the hardware into Device Support and Driver Support - which allowed Device Support to become generic for record types and drivers specific to handling different communication protocols.

And are there any papers regarding GTACS besides what is told in "EPICS:
A Control System Software Co-Development Success Story"
(https://epics.anl.gov/EpicsDocumentation/EpicsGeneral/epics_success.html)
and "EPICS Application Developer's Guide"
(https://epics.anl.gov/base/R3-14/12-docs/AppDevGuide.pdf)?

Many of these are listed below. I could not get access to some of them as they required institutional accounts. 

P.S. For me it looks like the initial design was VME-centric to allow
direct mapping of VAL fields to VME addresses so that reading and writing
becomes trivial, thus can be performed by "database engine" itself, while
Device Support and Driver Support were added later.  But I couldn't find
any confirmations or denials (neither any technical details of EPICS
structure/constitution nor rationale earlier than 1994).

VME and vxWorks were very convenient to use. A hardware engineer at Los Alamos, Bob Dingler. The 30 usec context switching and 2 usec interrupt latency made them good candidates. Still, we had Allen Bradley PLCs, motor controllers and GPIB/Serial devices that had to be integrated. Narrow APIs, Robust operation, deterministic execution and high performance were the primary design considerations.

It is important to mention that the technical issues had important but relatively small contributions to EPICS. The more relevant innovations were collaboration and open source development, which was enabled by Mike Thuot at LANL and Marty Knott at Argonne and fostered by many including Steve Lewis at LBL and Dave Gurd at the SSC, and then by Matthias Clausen at DESY Cryo, Carl (?) and Steve Hartman at Duke University, and Chip Watson / Karen White at JLab. The EPICS community is remarkable! Early on in the collaboration I was at a facility in Hawaii and had a technical issue that was answered by someone in Japan. When you attend an EPICS meeting or read through tech-talk, you find a lot of very dedicated people that have given time to help colleagues at other facilities in a wide variety of applications. That is the important part of EPICS. The community has been the most astounding accomplishment of EPCS. 

 

Some pre and early EPICS papers.

Recent Applications and Future Directions

Authors: Dalesio,,

Invited talk at the Particle Accelerator Conference, Chicago, Illinois, Date: June 18-22, 2001.

 

EPICS Directions to Accommodate Large Projects and Incorporate New Technology

Authors: Dalesio, Hill, et. al.,   

Talk at the International Conference on Accelerator and Large Experimental Physics Control Systems (ICALEPCS), Trieste, Italy, Oct. 4-8, 1999.

 

Data Archiving in EPICS

Authors: Kasemir, Dalesio,

ICALEPCS, Trieste, Italy, Oct. 4-8, 1999.

 

Integrating Commercially Acquired Subsystems

Author: Dalesio

ICALEPCS, Trieste, Italy, Oct. 4-8, 1999.

 

Data Archiving in Experimental Physics

Authors: Dalesio, Bickeley, et. al., ,

ICALEPCS, Beijing, China, Nov. 3-7, 1997.

 

The Success and Future of EPICS

Authors: Thuot, Dalesio, ,

Invited talk at the Linear Accelerator Conference, Geneva, Switzerland Aug. 26-30, 1996.

 

Distributed Software Development in the EPICS Collaboration

Authors: Dalesio, Kraimer, et. al., , i

Invited talk at the ICALEPCS, Chicago, Il. Oct. 30-Nov.3, 1995.

 

Control System Architecture: The Standard and Non-Standard Models

Authors: Thuot, Dalesio, ,

Invited talk at the Particle Accelerator Conference (PAC), May, 1993.

 

The Experimental Physics and Industrial Control System Architecture: Past, Present and Future

Dalesio, ,  

Invited talk at the ICALEPCS, Berlin, Germany October 1993.

 

The Sharing and Co-Development of the EPICS Software

Dalesio, et. al.,

ICALEPCS, Berlin, Germany October 1993.

 

A Performance Study of the Experimental and Industrial Control System

Author: Dalesio, et. al.,

ICALEPCS, Berlin, Germany October 1993.

 

A Distributed Timing System for Synchronizing Control and Data Correlation

Author: Dalesio, et. al.,

Neutral Particle Beam (NPB) Conference, Argonne, Ill. April 1992.

 

EPICS Architecture

Author: Dalesio, Kraimer, et. al.

Invited talk at the ICALEPCS, Tsukuba, Japan, 1991.

 

The Effects of the Toolkit Approach to Experimental Physics Control

Author: Dalesio, Thuot,

ICALEPCS, Tsukuba, Japan 1991.

 

A Virtual Control Panel Configuration Tool for the X-Windows System

Aurthor: Hill, Dalesio, et. al.

ICALEPCS, Tsukuba, Japan Nov. 1991.

 

Los Alamos Accelerator Control System: The Second Generation

Author: Dalesio, Hill, et.al.,

ICALEPCS, Vancouver, BC, Canada Nov. 1989.

 

The Los Alamos Accelerator Control System Database: A Generic Instrumentation Interface

Author: Dalesio,

Talk at the ICALEPCS, Vancouver, BC, Canada Nov. 1989.

 

 Run-Time Environment and Application Tools for the Ground Test Accelerator Control System

Authors: Kozubal, Dalesio, et. al.,

ICALEPCS, Vancouver, BC, Canada Nov. 1989.

 

 A State Notation Language for Automatic Control

Authors: Kozubal, Dalesio, et. al.,

ICALEPCS, Vancouver, BC, Canada Nov. 1989.

 

The Ground Test Accelerator Control System Database: Configuration, Run-time Operation, and Access

Author: Dalesio,

Particle Accelerator Conference, March 20-23, 1989, IEEE 0.1109/PAC.1989.72896

Thank you for indulging these ramblings. They represent only my perspective - there are a lot of different perspectives.

Bob

Attachment: PAC1989_1693.pdf
Description: Adobe PDF document