ACCOR PROJECT Strategy for Legacy Phase Out

1. ACCOR PROJECTStrategy for Legacy Phase Out M.Vanden Eynden on behalf ofall ACCOR project members and MCCs

2. Agenda • ACCOR Motivation for Legacy Phase Out • First Case : LynxOS-PowerPC platform • ACCOR Strategy for Legacy Phase Out • Key Objectives for LS1 • Conclusions ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

3. Motivation • ACCOR PROJET DEFINITION REPORT – June 2009 ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

4. Motivation • Several components beyond “end-of-life” • No support from industry (no spares) • Control system evolution almost impossible • Growing number of Operational Performance issues • Lack of clear eradication policy leads to • Multiplicity of solutions • Over-complexity and incompatibilities for renovation • CO resources locked in adaptative and corrective maintenance • Managing “end-of-life” is NOT a straightforward process • Corporate approach needed • Additional ACCOR complexity • Technology phase out goes together with new responsibility model at the front-end level ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

5. First Case • PowerPC Microprocessor Technology • No commercial contract anymore • Support limited to repairs only (if possible …) • Out of performance for new software solutions (CMW, Inca, more users, …) • Strategic decision to move towards Intel processors as common choice for cPCI, PICMG1.3 and VMEBus • LynxOS • Out of date, no industry support • Potential security issues • Prevent usage of modern software technologies (CMW, …) • Strategic decision to move towards Linux as THE multi-platform OS for cPCI, PICMG1.3 and VMEBus • Requires duplication of development and support for FESA, CMW, generic FESA classes, Device drivers, etc ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

6. Strategy for Phase out • Communicate the criticality of the situation • Assess the impact inside CO • Resources and planning for procurement of new solution • Impact on interfaces and other SW and HW components • Assess the operational impact • Number of systems concerned • Accelerators and Equipment groups concerned (induced effort, resources) • Define what “end-of-life” means • End of BE-CO operational support (no intervention in case of problem)? • Removal of operational equipment? • End of BE-CO procurement only? • Propose and explain alternate (new) solution or migration path • In case of agreement -> publish decision and plans • In case of disagreement • Re-asses alternate paths • Re-define roles and responsibilities • Re-iterate potential operation problems and their consequences (risk) ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

7. Strategy … but In Practice • BE-CO has to be more than Pro-active (MCCs) • The devil is in the details (“travail de fourmis”) • Identify ALL dependencies (“spaghetti effect”) • Do not underestimate priority clashes with other projects • Make compromises and offer alternate paths • Write strategy and objectives formally • Get agreement with partners ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

8. Strategyin a Nutshell ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

9. Key Objectives for LS1 • LynxOS-PPC platform eradication during LS1 for the following machines • LHC (> 200 FECs) • Move towards SLC6 64 bits (CO must be ready !) • BI : 125 FECs + 17 FESA Classes • BT : 7 FECs + SPS (6 FECs) • CO : 5 FECs • RF : 56 FECs • Interlocks : 24 to 48 FECs • LINAC3-LEIR Chain • OASIS • Timing System (TG8 eradication) • Non-CO FECs (about 20-25) • REX • Complete re-engineering ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

10. Key Objectives for LS1 • Major Implications of CO • Procurement process for • VMEBUs MEN SBCs • ELMA Crates • HW modules • Good synchronization with standard BE-CO HW modules roadmap (more on this later) • Linux 64 Bits kernel • Linux Device drivers • FESA 3.0 by MID-2012 (we are trying hard …) • Timing renovation (eradication of TG8s) • Generic FESA classes • Data Management tools for CCDB, Layout, Asset, CMM, including the migration to MTF • Extensive Installation coordination and Planning (> 400 FECs) ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

11. Key Objectives for LS1 ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

12. Key Objectives for LS1 • Standard BE-CO “hardware Catalog” roadmap • Severe operation problem in 2011 • Mandatory step before migration towards MTF • Other “time bombs” regarding spares, responsibilities, test beds, etc. • Complete cleaning done by Christine, Claude, Erik • 4 categories • Modules transferred to Eq.groups (+/- 25) • Modules which are obsolete and not used • Modules which are supported and up-to-date • Modules which are supported but under replacement (FMC, …) ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

13. Key Objectives for LS1 ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

14. Many other things for LS1 • Phasing out of Elonexplatform and replacement by Kontron PICMG1.3 • LHC : about 250 machines • Power converters • Cryogenics • Beam instrumentation • Beam Transfer • Etc. • Big KONTRON order launched 12/2011 (>400 systems) • Many other objectives • PS Continuous Transfer • AD High Tension Platform • OASIS (VXI replacement) • Etc . ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden

15. Conclusion • Going for a formal (EDMS) Legacy Phase Out Strategy by Accelerator is now a priority for ACCOR (in addition to all the other initiatives) • LS1 is a unique opportunity for ACCOR • Impact on ALL CO sections and key projects (Inca, FESA, etc.) • Very positive collaboration and agreement with Eq.Groups • Machine Controls Coordinators : Thank you ! • Key objective for LS1 : LynxOS-PPC phased out on LHC, REX, LINAC3 and LEIR • We are almost there … • LHC 100% • LINAC3-LEIR : see presentation of Marine • REX : see presentation of Frank ACCOR PROJECT – BE-CO-TC – M.Vanden Eynden