12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 1 Computing for CDF run II (2000 and beyond) (a work in progress !) SUMMARY Software.

Slides:



Advertisements
Similar presentations
Incontri di INFNET January 19, 1999 Impatto del Run II di CDF Stefano Belforte - INFN Pisa1 Reqs from CDF Run II on INFN computing infrastructures Usual.
Advertisements

Amber Boehnlein, FNAL D0 Computing Model and Plans Amber Boehnlein D0 Financial Committee November 18, 2002.
23/04/2008VLVnT08, Toulon, FR, April 2008, M. Stavrianakou, NESTOR-NOA 1 First thoughts for KM3Net on-shore data storage and distribution Facilities VLV.
O. Stézowski IPN Lyon AGATA Week September 2003 Legnaro Data Analysis – Team #3 ROOT as a framework for AGATA.
1: Operating Systems Overview
SM3121 Software Technology Mark Green School of Creative Media.
Trigger and online software Simon George & Reiner Hauser T/DAQ Phase 1 IDR.
G51FSE Version Control Naisan Benatar. Lecture 5 - Version Control 2 On today’s menu... The problems with lots of code and lots of people Version control.
The Project AH Computing. Functional Requirements  What the product must do!  Examples attractive welcome screen all options available as clickable.
Large scale data flow in local and GRID environment V.Kolosov, I.Korolko, S.Makarychev ITEP Moscow.
11 The Ultimate Upgrade Nicholas Garcia Bell Helicopter Textron.
Computer Programming My Home Page My Paper Job Description Computer programmers write, test, and maintain the detailed instructions, called programs,
Types of Operating System
L3 Filtering: status and plans D  Computing Review Meeting: 9 th May 2002 Terry Wyatt, on behalf of the L3 Algorithms group. For more details of current.
Command School On Task In Touch Online Software for Schools Developed by Schools.
The D0 Monte Carlo Challenge Gregory E. Graham University of Maryland (for the D0 Collaboration) February 8, 2000 CHEP 2000.
SLIR Computer Lab: Orientation and Training December 16, 1998.
14th IEEE-NPSS Real Time Conference 2005, 8 June Stockholm.
Online Systems Status Review of requirements System configuration Current acquisitions Next steps... Upgrade Meeting 4-Sep-1997 Stu Fuess.
Introduction and Overview Questions answered in this lecture: What is an operating system? How have operating systems evolved? Why study operating systems?
Test Of Distributed Data Quality Monitoring Of CMS Tracker Dataset H->ZZ->2e2mu with PileUp - 10,000 events ( ~ 50,000 hits for events) The monitoring.
Planning and Designing Server Virtualisation.
Remote Production and Regional Analysis Centers Iain Bertram 24 May 2002 Draft 1 Lancaster University.
Computing and the Web Operating Systems. Overview n What is an Operating System n Booting the Computer n User Interfaces n Files and File Management n.
Royal Latin School. Spec Coverage: a) Explain the advantages of networking stand-alone computers into a local area network e) Describe the differences.
03/27/2003CHEP20031 Remote Operation of a Monte Carlo Production Farm Using Globus Dirk Hufnagel, Teela Pulliam, Thomas Allmendinger, Klaus Honscheid (Ohio.
Farm Management D. Andreotti 1), A. Crescente 2), A. Dorigo 2), F. Galeazzi 2), M. Marzolla 3), M. Morandin 2), F.
+ discussion in Software WG: Monte Carlo production on the Grid + discussion in TDAQ WG: Dedicated server for online services + experts meeting (Thusday.
Cloud Computing Characteristics A service provided by large internet-based specialised data centres that offers storage, processing and computer resources.
D0 SAM – status and needs Plagarized from: D0 Experiment SAM Project Fermilab Computing Division.
3rd June 2004 CDF Grid SAM:Metadata and Middleware Components Mòrag Burgon-Lyon University of Glasgow.
System Security Chapter no 16. Computer Security Computer security is concerned with taking care of hardware, Software and data The cost of creating data.
MiniBooNE Computing Description: Support MiniBooNE online and offline computing by coordinating the use of, and occasionally managing, CD resources. Participants:
Block1 Wrapping Your Nugget Around Distributed Processing.
Jan. 17, 2002DØRAM Proposal DØRACE Meeting, Jae Yu 1 Proposal for a DØ Remote Analysis Model (DØRAM) IntroductionIntroduction Remote Analysis Station ArchitectureRemote.
CS 390 Unix Programming Summer Unix Programming - CS 3902 Course Details Online Information Please check.
Plans for Trigger Software Validation During Running Trigger Data Quality Assurance Workshop May 6, 2008 Ricardo Gonçalo, David Strom.
Personal Computer - Stand- Alone Database  Database (or files) reside on a PC - on the hard disk.  Applications run on the same PC and directly access.
CDF Offline Production Farms Stephen Wolbers for the CDF Production Farms Group May 30, 2001.
21 st October 2002BaBar Computing – Stephen J. Gowdy 1 Of 25 BaBar Computing Stephen J. Gowdy BaBar Computing Coordinator SLAC 21 st October 2002 Second.
9 February 2000CHEP2000 Paper 3681 CDF Data Handling: Resource Management and Tests E.Buckley-Geer, S.Lammel, F.Ratnikov, T.Watts Hardware and Resources.
5 May 98 1 Jürgen Knobloch Computing Planning for ATLAS ATLAS Software Week 5 May 1998 Jürgen Knobloch Slides also on:
Introduction to WebFX. Summary of Last Lectures n Introduction to computers n Computer hardware and software –computer microcomputer minicomputer mainframe.
Febryary 10, 1999Stefano Belforte - INFN Trieste1 CDF Run II Computing Workshop. A user’s perspective Stefano Belforte INFN - Trieste.
EGEE is a project funded by the European Union under contract IST HEP Use Cases for Grid Computing J. A. Templon Undecided (NIKHEF) Grid Tutorial,
240-Current Research Easily Extensible Systems, Octave, Input Formats, SOA.
Test Results of the EuroStore Mass Storage System Ingo Augustin CERNIT-PDP/DM Padova.
Analysis trains – Status & experience from operation Mihaela Gheata.
CDMS Computing Project Don Holmgren Other FNAL project members (all PPD): Project Manager: Dan Bauer Electronics: Mike Crisler Analysis: Erik Ramberg Engineering:
Online Monitoring for the CDF Run II Experiment T.Arisawa, D.Hirschbuehl, K.Ikado, K.Maeshima, H.Stadie, G.Veramendi, W.Wagner, H.Wenzel, M.Worcester MAR.
Status of the new NA60 “cluster” Objectives, implementation and utilization NA60 weekly meetings Pedro Martins 03/03/2005.
UTA MC Production Farm & Grid Computing Activities Jae Yu UT Arlington DØRACE Workshop Feb. 12, 2002 UTA DØMC Farm MCFARM Job control and packaging software.
Computing R&D and Milestones LHCb Plenary June 18th, 1998 These slides are on WWW at:
Feb. 14, 2002DØRAM Proposal DØ IB Meeting, Jae Yu 1 Proposal for a DØ Remote Analysis Model (DØRAM) Introduction Partial Workshop Results DØRAM Architecture.
CD FY09 Tactical Plan Status FY09 Tactical Plan Status Report for Neutrino Program (MINOS, MINERvA, General) Margaret Votava April 21, 2009 Tactical plan.
Upgrade Software University and INFN Catania Upgrade Software Alessia Tricomi University and INFN Catania CMS Trigger Workshop CERN, 23 July 2009.
Feb. 13, 2002DØRAM Proposal DØCPB Meeting, Jae Yu 1 Proposal for a DØ Remote Analysis Model (DØRAM) IntroductionIntroduction Partial Workshop ResultsPartial.
A UK Computing Facility John Gordon RAL October ‘99HEPiX Fall ‘99 Data Size Event Rate 10 9 events/year Storage Requirements (real & simulated data)
D0 Farms 1 D0 Run II Farms M. Diesburg, B.Alcorn, J.Bakken, R. Brock,T.Dawson, D.Fagan, J.Fromm, K.Genser, L.Giacchetti, D.Holmgren, T.Jones, T.Levshina,
Hans Wenzel CDF CAF meeting October 18 th -19 th CMS Computing at FNAL Hans Wenzel Fermilab  Introduction  CMS: What's on the floor, How we got.
Apr. 25, 2002Why DØRAC? DØRAC FTFM, Jae Yu 1 What do we want DØ Regional Analysis Centers (DØRAC) do? Why do we need a DØRAC? What do we want a DØRAC do?
Analysis Model Zhengyun You University of California Irvine Mu2e Computing Review March 5-6, 2015 Mu2e-doc-5227.
CDF ICRB Meeting January 24, 2002 Italy Analysis Plans Stefano Belforte - INFN Trieste1 Strategy and present hardware Combine scattered Italian institutions.
SuperB and its computing requirements
Experiences and Outlook Data Preservation and Long Term Analysis
The ZEUS Event Store An object-oriented tag database for physics analysis Adrian Fox-Murphy, DESY CHEP2000, Padova.
OffLine Physics Computing
Get your ETL flow under statistical process control
Using an Object Oriented Database to Store BaBar's Terabytes
Presentation transcript:

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 1 Computing for CDF run II (2000 and beyond) (a work in progress !) SUMMARY Software engineering: –Object Oriented migration –Languages Data –size –storage/handling Analysis in Pisa –data size –computing model –hardware needs Network –analysis needs –on-line monitoring needs Commercial software A “keep your feet on the ground” approach

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 2 Software Engineering I CDF will evolve to OO approach, large emphasis on capitalizing on past code and experience: –mixed language environment C++ basic choice for Run II offline support FORTRAN as well (expect it will die sometime) JAVA in online only so far (as well as C) –wrap old code in C++ and run in new environment –data access still allowed via BANKS, OO is encouraged but not required software engineering tools left to experts so far, keep design simple, concentrate on physics, a lot of time wasted on class charts on thin air, real stuff starting now learning curve on new paradigm just started, 90% of new code is in C++, but capability to run together with old Fortran code to check new algorithms and build right away complete package was vital

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 3 Software Engineering II Simulation, reconstruction and analysis code as user modules embedded in a general framework that provides: –data I/O, user interfaces, dynamic module invocation, data-base communication etc. etc. framework (A_C++) developed together with BaBar, developed by professionals user modules written by physicists, OO/C++ experts to provide assistance, some more complex modules maybe written by experts,but most written by “everybody” from old timers (Fortran IV addicts) to young students: physics competence first ! It is a very large project, startup was difficult, is one major concern of the collaboration, now you know as much as I do.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 4 Compilers and Platforms It it would only be the new code, it would be a breeze. But CDF wants to carry on the good old stuff. Lesson from the past: as standard as possible, only ANSI compliant C++ compilers. KAI (commercial one) only supported one so far. Platform support: needs driven: –SGI: is just the machine we have in FNAL –Linux: is what we will use for Production and Level 3 –Digital Unix: just because we (Pisa and many other universities) asked for it, and we (Pisa) are working on it (with Padova) –Solaris: is talked to be inevitable, but still has to come –AIX: likely disappearing (no-one wants it) The party line is to do things so that platform migration will be easy all in all we already survived VMS  ACP  AMDAHL  UNIX ! “Usual” FNAL story: everything is manpower limited.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 5 Code Distribution Code management via CVS Release management via Soft_Rel_Tool (shared with BaBar) So far FNAL only supports distribution as “packaged products” via laboratory owns ups/upd methods No AFS Code distributed as tar files, may need local rebuilding CDF will need local system manager help for this !

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 6 Data Storage Not an easy decision, past was YBOS sequential files… would not scale easily to run II data size, but worked  keep it for initial debugging of detector and code and as fallback. Then choose among: –Objectivity, home-grown LWDB, old-fashioned sequential files, ROOT. we just had: –1 year debate, 3 months technical review, 2 day workshop, recommendation came out 5 days ago, still wait definitive management ruling. –recommendation: ROOT …. But compatible with YBOS sequential files and leave door open for objectivity in the future.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 7 Data Size Data logging: 75 Hz for 2 years (2 inverse fb) Raw event size: 250 KB Add analysis data and data overlap: 520 KB Reduced analysis data set (PAD) = 60 KB/event hope to get  30KB with ROOT compression Overall data set for Run II: 1 PB = 1000 TB Overall PAD size: 160 TB FNAL data storage (all data are at FNAL, at least!) – 30 TB disk (25TB PADs + 10 TB robot cache) –400TB tape robot (a guess so far) –rest on shelf tapes. 200 physicists doing analysis at FNAL, about 20 at Pisa, still simple scaling don’t work (can’t have 3 TB disk + 40 TB robot in Pisa, data can not be parted among physicists).

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 8 Data Analysis in Pisa I Scenario 1: copy here the PAD’s we are interested in: W/Z : 2 million events = 60 GB (gets top as well !) J/psi : 80 million events = 2.4 TB B->pipi: 1 Hz for 1 year = 30 million events = 1 TB –maybe less inclusive, maybe less compression : 3 TB total –need: 3TB robot, O(500GB) disk, about 3 times na48farm, buy it in … same cost ? –What about SUSY, jets.. ? Scenario 2: keep PADs in FNAL, copy here n-tuples: wild guess: will need 10 GB per physicist adding code, common stuff etc. 300GB disk, plus HSM for backups etc, size as present na48farm BUT : n-tuple refresh from FNAL ! One set a week means 10GB * 20 users = 200GB/week = 500 KB/sec Fnal-to-Pisa

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 9 Data Analysis in Pisa II Scenario 3: keep everything in FNAL. just use X-terminals in Pisa and run everything through the network need as good a line as from my home 28.8K modem = 4KB/sec/user = 80 KB/sec to FNAL past attempts always failed, even now file transfer to/from FNAL is O(10KB/sec) (Netscape e.g.) but interactive shells are already almost impossible to use. Scenario 4: keep local copy only of currently used n-tuple is like 2 (need local power comparable to na48farm) but limit refresh rate to 1GB/week/user (already can hear people screaming for more!) still need 20GB/week = 50 KB/sec guaranteed bandwidth.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 10 Data Analysis in Pisa: The Final Answer We will have to try, can’t pick the right approach before collaboration has finalized data handling and distribution tools, and analysis topics have been pinpointed We will try everything, user pressure will drive Needs will be dominated by physics output (maybe we find SUSY in 3-lepton samples and everybody looks at this small data set…) We will exploit local computing as much as possible to reduce network load (likely bottleneck, as it always has been) Still will need to access FNAL PADs to produce data sets to copy to Pisa. If network is no good will use taped (expensive though!). But we desperately need guaranteed bandwidth for interactive work If can not log in FNAL, no way to do most analysis here, only use “dead” data sets: no express data, no hot topics, just late sidelines… the good old way: take the plane and go to FNAL.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 11 Analysis in Pisa: My personal bottom line The CDF group in Pisa will put together a local computing facility which will not cost more than what we have spend e.g. for Run I analysis or na48farm: one central server with 3TB robotic tapes and O(500) GB disks, O(10) top of the line PC’s running Linux with 10GB local disk each. We will be able to really do competitive analysis on hot topics only if INFN can provide really good network for interactive response. –10 min to download a few Mbytes postscript document: OK –10 sec to move the cursor on a remote editor: UNACCEPTABLE

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 12 Pisa: Mainframe or Cluster ? Need mainframe for robot, file server etc. PC farm supported via Linux, will use as needed Will mostly depend on hardware cost at time of buy (2 years from now or more) past experience with scattered hardware taught that it must be avoided hardware on desk is no good »hot »noisy »damaged hardware bought a bit at a time is a mess distributed computing is a maintenance nightmare Anyhow, we will probably have both: efficiency and easy of management call for central facility (at present also cost effective), users will have PC’s on their desks anyhow.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 13 Italy: global center or everyone by himself ? Data are just too much, can not bring to Italy all that we need and cut the line with FNAL. Thus we need a good link (see above) from each INFN site to FNAL anyhow. Given a good access to FNAL resources for remote job entry at least, local requirements “per user” are small, sites with less then 5 physicists working full time on analysis at any given time may just do with a few PC-like workstations. At present we do not see a need, nor an advantage, in concentrating computing from more INFN sites in just one. For example in the past different INFN sites picked different analysis topics (very sensible), then is better to keep everybody’s data near the user. Global center may all in all enhance network needs, I would like to see the lesser requirements satisfied first !

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 14 Network needs Offline covered before: need guaranteed interactive work, will probably be OK if heavy graphics is slow. Will transfer large data sets via tapes. Online: a new front. –1. Silicon Vertex Tracker: a crucial part of Level 2 trigger, needs a lot of online monitor, care, debugging: experts will be in Pisa –2. Internal Silicon Layers: our largest contribution to new detector, want to keep ownership after installation, needs on-line assistance from Pisa as well. –3. Remote Control Room: if we can do shifts from Pisa, we save on travel –3. Is about saving money, 1 and 2 about getting proper credit and physics rewards from present large efforts: keep ownership of our detectors ! –NEED: 2 or 3 X-terminals that work as extensions of FNAL LAN

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 15 Extending FNAL LAN Needed for online monitoring and debugging and remote control room Do not need full 1MB/sec Ethernet, need tests but probably 200 KB/sec is good enough. Do not need it at all times (problem in FNAL, phone call to expert in PISA, get the line for a few hours) It has to be a guaranteed connection, something to rely on, if it fails a few times, the collaboration will request expert on site Can “you” do it ? Can I have a dedicated slice of the net kept aside for just my terminal ? Do we have to buy 64K ISDN by the minute as for videoconference ? Lot of work to do, but need agreement on direction to move.

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 16 Commercial Software Back to the beginning: CDF is a dynosaur slowly evolving from Run I environment, no jumping. DI3000 in Run I: bad experience. Present status (the way I read it): it is being introduced, but slowly, and most collaboration (rightfully) pull the brakes, no special enthusiasm. Software packages in use –general ROOT cvs ZOOM (FNAL C++ Cernlib replacement) KAI C++ compiler mSQL (proposed) –just for professionals memory analysers (Sniff++ etc.) OO development tools

12 Maggio 1998Computing for CDF run II Sefano Belforte - INFN Pisa 17 Conclusion The picture is starting to clear out only now –if this worries you, at least I am not alone Can not show a real neat plan, still... Will not need much computing harware before data, we are confident we can handle that pretty much as last run, need local help for system management and code distribution. We have higher goals this time: –competitive top of the line physics –keep ownership of detectors we build For this we have new needs: –network ! –network ! ! –network ! ! !