and if we d id it all again…?

13
+ And if We Did it All Again…? G. Lehmann Miotto

Upload: eyad

Post on 15-Feb-2016

31 views

Category:

Documents


0 download

DESCRIPTION

And if We D id it All Again…?. G. Lehmann Miotto. DAQ@LHC: What did I learn?. We’ve been extremely successful at solving the same problems in 4++ different ways For data-flow architectures one may still argue that differences were driven by Different experiment requirements - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: And if We  D id it All Again…?

+

And if We Did it All Again…?G. Lehmann Miotto

Page 2: And if We  D id it All Again…?

DAQ@LHC Workshop

2+DAQ@LHC: What did I learn? We’ve been extremely successful at solving

the same problems in 4++ different ways For data-flow architectures one may still argue

that differences were driven by Different experiment requirements Different risk assessment/perception of experts

For online software (control, configuration, monitoring) it is really difficult to find a technical explanation Different people naturally chose different options

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 3: And if We  D id it All Again…?

DAQ@LHC Workshop

3+Today’s Online Systems4 well performing systemsstill improving/evolving

Ecumenical Institute - Chateau Bossey, March 14th 2013

• Evolving technologies• Keep up• Explore and track• Profit

• Increase automation & usability• Note that effort is needed also to

KEEP the present perfection level!

Page 4: And if We  D id it All Again…?

DAQ@LHC Workshop

4+How Did We Get Here? A challenging set of problems The lack of early existing solutions caused

research work to start independently, in parallel (even within experiments)

Initially plenty of human resources (read eager experts) organized in separate communities R&D is what we love! No expert would willingly give up his solution

Forcing convergence seemed counterproductive

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 5: And if We  D id it All Again…?

DAQ@LHC Workshop

5+DCS Took a Different Path Long initial discussions to agree on a (third

party) solution Customized to experiments’ needs Integrated with in-house developed components

Coordinated outside individual experiments (JCOP)

We have working detector control systems for all experiments=> also this seems a valid approach

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 6: And if We  D id it All Again…?

DAQ@LHC Workshop

6+What if We Did It All Again Today?Still a challenging set of problemsR&D is still what we loveWe are organized in separate

communitiesExcept if there was a strong managerial

decision…

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 7: And if We  D id it All Again…?

DAQ@LHC Workshop

7+But… We should keep in mind that we are working on

an experiment lifetime of 20 – 30 more years We are not even half way through With how many (4 * re-implementations) should we

count?

Will we maintain a similar level of resources in the long term? Investment from institutes Interest of new people

Therefore, it is not completely academic to ask ourselves…

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 8: And if We  D id it All Again…?

DAQ@LHC Workshop

8+Where Do We Go from Here?Possible options:

a) Continue as today

b) Pick up the best from each system / service / component and “merge” into a common solution

c) Evaluate a new, common approach

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 9: And if We  D id it All Again…?

DAQ@LHC Workshop

9+a) Continue as Today After all…. nobody would expect religions to merge after

an ecumenical gathering It may be simplest to continue improving our own systems Now that we know better what the other experiments are

doing We may ask for expert opinions before embarking onto the

next software evolution We may decide to pick up some of the ideas that were

presented for other experiments and implement them in our system

We may choose to use common technologies in specific areas We may even launch some common projects at the periphery

of our systems

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 10: And if We  D id it All Again…?

DAQ@LHC Workshop

10+b) Select & Merge May seem a viable approach, but…

Ecumenical Institute - Chateau Bossey, March 14th 2013

• We’d face the problem of fitting a square peg into a round hole…

• A patched solution to nobody’s satisfaction

Page 11: And if We  D id it All Again…?

DAQ@LHC Workshop

11+c) Go for a Common Approach A major effort in social

engineering/team building

If we want to have a chance of success Take time to look at how other environments have tackled

the problem Computer centers Distributed/networked control systems

Don’t exclude a priori to go the SCADA/OPC UA way

At the very least, agree on and choose a common middleware The only way to keep the door open to “merging” later on

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 12: And if We  D id it All Again…?

DAQ@LHC Workshop

12+Summary So far we have been very successful with our

independent solutions The long experiment lifetime demands that we ask

ourselves how to proceed We will re-design & re-implement our systems a few

more times This workshop has offered the possibility to establish

communication channels Up to us to decide to what extent it’s better to work

independently or together

Ecumenical Institute - Chateau Bossey, March 14th 2013

Page 13: And if We  D id it All Again…?

+

A Big ThanksTo Clara, Vasco, Luciano for the preparatory workshop discussions and to ALL of you for this nice ecumenical workshop!