Controls Strategy- Lehman Review – Feb 2004 1 Controls Strategy Joel N.Butler Talk to Department...
-
Upload
philip-walters -
Category
Documents
-
view
217 -
download
0
description
Transcript of Controls Strategy- Lehman Review – Feb 2004 1 Controls Strategy Joel N.Butler Talk to Department...
Controls Strategy- Lehman Review – Feb 2004 1
Controls StrategyControls Strategy
Joel N.ButlerTalk to
Department of Energy Assessment of the Run II Luminosity Plan at the Fermilab Tevatron
Accelerator Integration SessionFeb. 24, 2004
Controls Strategy- Lehman Review – Feb 2004 2
What is the Concern?What is the Concern? The Control System works well but could work better
and contains obsolete components The character of the Run 2 upgrades puts a lot of pressure
on the system. We are also for the first time running Fixed Target and Collider simultaneously. The system must not only control and monitor the complex but must provide data acquisition and archiving, support data analysis and trending, provide more automated control and feedback, and support complex accelerator studies and calculations
It must support/facilitate development of new applications It must operate efficiently and be easy to maintain It must continue to do these tasks for many years
There are needs that are not optimally addressed by the current system, so process improvement is worth pursuing.We are undertaking a comprehensive review of the system
Controls Strategy- Lehman Review – Feb 2004 3
Goal of the Current ReviewGoal of the Current Review We want to understand the requirements
of the system and support departments in order to optimize the long term strategy
We aren’t trying to break things We aren’t trying to make extra work for
people We are very aware of the need to
maintain the ability to operate even while we undertake improvements.
Controls Strategy- Lehman Review – Feb 2004 4
ConstraintsConstraints No long shutdowns to replace large blocks of functionality
Precludes fundamental redesign of the system Parts of new system must be plug compatible and
communicate with old Must be able to easily switch new things in and back out
and/or run them in parallel No big pile of money to replace everything
Base on inexpensive commodity computing equipment Many small PCs/unix boxes rather than few large systems
No big influx of new people for a parallel effort Work must be done by existing lab personnel.
No long break to learn a new system New languages, tools etc. should have modest learning curve A lot of code written by non-Computing Professionals System department people are busy with accelerators
Controls Strategy- Lehman Review – Feb 2004 5
Control System OverviewControl System Overview
Applications
CentralServices
Front-Ends
Applications Applications
Front-Ends Front-Ends
CentralServices
…..
…..
…..
Field Hardware
ethernet
field bus
…..
Controls Strategy- Lehman Review – Feb 2004 6
Overview (VMS)Overview (VMS)
VAXes
Computer Room
Or elsewhere
The VAX 4090’s
are obsolete and
unsupported
Front Ends
(anywhere)………………
Datapool
Manager
Datapool
Manager………..
Applications Applications ………………
Front-End Front-End
Shared Memory
ACNET/Ethernet
Ethernet
X WindowsDisplay(PC)
X WindowsDisplay(PC)
…
Mature libraries and services to facilitate applications development. System is written in C (assembler) and relies heavily of VMS operating system
>500 “Console Applications”, those with user interfaces, written in C and FORTRAN. These do much of the work in support of operations
Controls Strategy- Lehman Review – Feb 2004 7
Overview (Java)Overview (Java)
Sun Netras
(computer room)
PCs
(anywhere)
Front Ends
(anywhere)
DAE DAE
………………
(16 consolidators)
DSE----------------OAC/Servlet
DSE----------------OAC/Servlet
………..
Application Web Application ………………
Front-End Front-End
RMI
ACNET
ACNET
http
This system could be a complete replacement on a more modern, operating system independent, platform. It is used for many non-interactive task (OACs), for SDA, Autotune for MiniBooNe and Ecool, and a few interactive applications.
From here, same protocoland database tables as VAXs
Controls Strategy- Lehman Review – Feb 2004 8
IssuesIssues VAXes
obsolete and constraining• Too little CPU power, memory, disk storage• There are many workarounds to solve problems
Have the most complete application support, but not easily portable
Familiar environment. AD people know how to program in C and FORTRAN. The applications are mostly portable if the new system’s infrastructure was made to “look like the VAX”.
JAVA Modern computers and programming language with
plenty of resources and easy expansion capability Incomplete support for applications Unfamiliar programming paradigm and language.
“Porting”in the usual sense of the word is not possible
Migration was assumed to involve a rewrite Because of the nature of Java, has some
performance and tuning issues as well
Controls Strategy- Lehman Review – Feb 2004 9
Approach: Controls Working GroupApproach: Controls Working Group We need to find out
What we are running now and what problems there are
What we would like to be running now if only we had the resources (especially programmers)
What new applications we need to develop in the future and what resources it will take to acquire them and run them to full advantage
What specialized and embedded systems we support, how they are supported and how they integrate into operations
What specialized systems and capabilities we imagine supporting in the futureWe set up a group consisting of two representatives from
each department and a few at large members to answer thesequestions – a.k.a the Controls Working Group (CWG).
Controls Strategy- Lehman Review – Feb 2004 10
Schedule of CWG TalksSchedule of CWG TalksOct. 29 Controls (Patrick) Booster(Kasper)
Nov. 5 Tevatron (Moore)
Nov. 12 Controls (Patrick)
Nov. 19 PBAR (Derwent) PBAR (Peterson)
Dec. 10 Controls (Hendricks_ Controls(Hendricks)
Dec. 17 Main Injector (Capista) Taking stock
Jan. 7 External Beams (Lucas) RF (Chase) ES&H (Zimmerman)
Jan. 14 Electrical (Martin) CR YO (Norris)
Jan. 21 Instrumentation (Voy) Labview devices(Lorman)
Inst/Introduction(Pordes)
Jan. 28 Recycler(Gattuso)
Feb. 4 ECOOL (Warner) SDA (Slaughter)
Feb. 11 Linac(Allen) Operations(Rohde)
Mar. 3 Operations (Kissel)
Controls Strategy- Lehman Review – Feb 2004 11
Future NeedsFuture Needs A major challenge for the Booster is to achieve
and MAINTAIN peak performance …
Day to day and even hour to hour scatter is large!
Maximum p/hr allowed by BLM’s
Protons/pulse
Controls Strategy- Lehman Review – Feb 2004 12
Monitoring ChangesMonitoring Changes Need to continuously monitor and record
changes in the machine’s behavior Complicated by the rapid cycling nature of the
Booster Would like to monitor hundreds of devices on
minute time scales but are limited by bandwidth issues
… lots of potential for improvement.. Can provide alarm capability for snapshot
devices Change analyses Auto tune
Controls Strategy- Lehman Review – Feb 2004 13
Key General ThemesKey General Themes MORE BANDWIDTH! MORE DATA STORAGE!
We are storing more and more stuff in the data-loggers
We would like it to stay around for a much longer time (years not months)
MORE AUTOMATION “It’s the 21st Century. Why not think more about
smart systems.”• Automated troubleshooting• Automated beam tuning
These, together with the desire for a more convenient andeasier application development environment were thekey general themes we heard from the CWG
Controls Strategy- Lehman Review – Feb 2004 14
Migration to JAVAMigration to JAVA We did a survey of VMS Console Applications:
Not counting obsolete or redundant programs Estimating what could be gained by application consolidation
either within departments or across departments We were pleased to learn that there was a fairly high degree of
consolidation, even across departments We have now established that there is a base of VMS console
applications that is greater than 500 programs Most people claimed that they did not have time to learn a new
programming language. Nor did they have time to “rewrite” their programs in this new language
There has not been an aggressive program to teach people JAVA nor were there good tutorials anchored in the actual tasks people were trying to do
There are pieces of the JAVA support infrastructure missing. Many needs are too pressing to wait for JAVA to catch up so even some new APPS are written for the VAX
Controls Strategy- Lehman Review – Feb 2004 15
The Emergence of a New ProposalThe Emergence of a New Proposal The Controls Group has proposed to attempt to
“port” the VAX/VMS CONSOLE INFRASTRUCTURE to a LINUX/INTEL platform.
Then, most of the Console Applications should simply port
The much more modern, powerful processors should solve a number of problems and applications can be extended in a familiar language.
This proposal makes sense ONLY if the Console Infrastructure can be rewritten QUICKLY The goal is for enough work to be done to port some
applications by the summer and To have most of the Console Applications migrated by
Oct 2005 These are just goals at present – planning/scheduling is
just beginning There will be a review of the proposal, including its
implementation, in March. A charge has been written.
Controls Strategy- Lehman Review – Feb 2004 16
What Happens to JAVA?What Happens to JAVA? The JAVA environment has proven to be very
useful and many modern Accelerator controls systems are based on it. In the end, it is believed to provide the best path to an easier application development environment
We will try to overcome some of the”acceptance” problems by Developing JAVA tutorials Producing examples of how to write common
types of applications Completing missing infrastructure Improving documentation Establishing a JAVA applications writer group
Controls Strategy- Lehman Review – Feb 2004 17
How will it come together?How will it come together?
The JAVA system already runs on LINUX. Then, we will have one platform that can run C and JAVA
Somewhat farther down the road, one could see that a single set of common services could support either C or JAVA-based applications
The Controls Department will be supporting at least two systems for some time. The Department will have to devote resources to develop the infrastructure and some of the applications support will have to revert to physicists and others This requires the Department to play more of a leadership role and less of a support role. They will have to guide people to use the most efficient and appropriate tools for their applications. This will require management and the Department to work together to get the correct result. Priorities and rules will need to be set and enforced.
Controls Strategy- Lehman Review – Feb 2004 18
Other Lessons from the CWGOther Lessons from the CWG Communication on controls issues is not great (for
many good reasons) Consequently, “best practices” are not uniformly
applied across the Division Knowledge of the capabilities and support available
from the current system is not good Needs and problems are not well-communicated Priorities and schedules are not always well-
established We will continue the CWG (perhaps biweekly) and
turn it into a forum for addressing controls issues across all departments. Examples are hardware initialization Save and Restore Writing of non-interactive tasks (OACS)
Controls Strategy- Lehman Review – Feb 2004 19
• The Control System is not broke, but can work better to serve Fermilab’s goals.
• We will work towards improving it • We hope that it will lead to easier operations and
maintenance, more consistency, more automation, and more efficient studies and in that way contribute to the Run 2 luminosity improvements.
ConclusionConclusion
Controls Strategy- Lehman Review – Feb 2004 20
ConclusionConclusion
The Control System is not broke, but can work better to serve Fermilab’s goals. We will work towards improving it. We hope that it will lead to easier operations and maintenance, more consistency, more automation, and more efficient studies and in that way contribute to the Run 2 luminosity improvements.