1. This material is based upon work supported by the U.S. Department of Energy Office of Science under Cooperative Agreement DE-SC0000661. Michigan State University designs and establishes FRIB as a DOE Office of Science National User Facility in support of the mission of the Office of Nuclear Physics. Vasu Vuppala FRIB Data Services - Design

2.  Information Architecture  Design Principles  Domains  Components  Hierarchies  Lattice - Models  EPICS  Cabling  Security  Challenges Overview V. Vuppala, Controls DB Meeting, Slide 2

3.  Database (DB): An organized collection of data  Database Management System (DBMS): Software system that manages databases  Database Model: Structure of the data in a database Hierarchical, Network, Relational, ER, Object, Dimensional etc  Relational Model: Based on mathematical relation (set theory)  Relational DBMS: DBMS based on Relational Model Ingres, Oracle, DB2, MySQL, Postgres etc Terminology V. Vuppala, Controls DB Meeting, Slide 3

4. Information Architecture, Slide 4V. Vuppala, Controls DB Meeting  Application layer Operator interfaces High-level applications Libraries  Service layer Access to data Programming Interface  Data layer Managed data No direct access RDB, Files, No-SQL etc

5. Distributed Components, Slide 5V. Vuppala, Controls DB Meeting  Distributed across the network Data Services Applications  Scalable

6.  Keep the schema as simple as possible.  Do not model every small piece of the facility.  The operational database must be in sync, or very close to, the real facility. So it should be easy to update the database.  Updates to the database should be controlled but not overly cumbersome.  The schema should be generic, to an extent, so that other labs can use it. However, it should not be at an inordinate cost.  It should be possible, for other labs, to use parts of the schema, possibly by having a set of loosely-coupled subschemas or “domains”. DB Design Principles V. Vuppala, Controls DB Meeting, Slide 6

7. Schema Domains V. Vuppala, Controls DB Meeting, Slide 7

8. FRIB Configuration, Slide 8V. Vuppala, Controls DB Meeting

9. FRIB Configuration – LS1, Slide 9V. Vuppala, Controls DB Meeting

10. Components V. Vuppala, Controls DB Meeting, Slide 10  Physical Component: This represents physical entities; things that exist in the real world. Example: the cavity with part number T30802-MDE-0008 that was manufactured by Jefferson Lab is a physical-component.  Conceptual Component or Component Type This represents generic entities or concepts and their designs. Example: horizontal electrostatic dipole, beam current monitor, RFQ, and RF cavity  Configuration or Logical Component It represents the entities that exist on the blueprint or configuration of the FRIB facility. Example: the cavity LS1_CA01:CAV2_D1101

11. Components: Example V. Vuppala, Controls DB Meeting, Slide 11

12. Components Schema V. Vuppala, Controls DB Meeting, Slide 12

13. Configuration Components V. Vuppala, Controls DB Meeting, Slide 13

14. Configuration Hierarchy V. Vuppala, Controls DB Meeting, Slide 14

15. Control Hierarchy V. Vuppala, Controls DB Meeting, Slide 15

16. Configuration Component Relations V. Vuppala, Controls DB Meeting, Slide 16

17. Conceptual Components V. Vuppala, Controls DB Meeting, Slide 17

18. Q1Q2Q3 Location in driver linacLS1, FS1, FS2, LS3, BDSBDS Effective Length (m)0.250.260.40 Full Aperture, A (min. pipe ID, m) 0.050.075 Quadrupole Gradient, B’0 (T/m)3530 Approx. Quadrupole Pole Tip Field (T)0.875*1.2 Quadrupole good field region, 2 R0 = 0.8 A (m) 0.040.06 Quadrupole field uniformity dB’/B’0 within radius R0, over operating range ≤ ±5e-3≤ ±7e-3 Operating range, as fraction of B’070-100% Quadrupole Specifications V. Vuppala, Controls DB Meeting, Slide 18

19.  Set of Tables A Table (or set of tables) for Each Conceptual Component Pros: Efficient Storage and Performance, Relational Cons: Limited Schema Changes Issues; Unsupported Data Types (ranges, lists etc) Implementation 1 V. Vuppala, Controls DB Meeting, Slide 19

20.  Set of Properties A Table of Attributes Table for Conceptual Components Pros: Flexible (Schema Changes) Cons: Efficiency, Performance, Complex SQL Issues: Storage of Different Data Types Implementation 2 V. Vuppala, Controls DB Meeting, Slide 20

21. Physical Component V. Vuppala, Controls DB Meeting, Slide 21

22. Implementation V. Vuppala, Controls DB Meeting, Slide 22

23.  What are all the Horizontal Electrostatic Quads?  Who is the manufacturer of ‘LS1_CA01:CAV2_D1101’?  Where is T30802-MDE-0008 installed?  What are all the components in LINAC Segment 1?  What is CAD drawing for LS1_CA01:CAV2_D0910?  When was LS1_CA01:CAV3_D1101 calibrated?  What are the PVs associated with component with serial number T30802-MDE-008?  Is LS1_CA01:PS1_D0910 turned on?  Give the traveler information associated with T30802-MDE-008. Data Service V. Vuppala, Controls DB Meeting, Slide 23

24.  Components Conceptual (Component Type) Configuration (Logical) Physical  Design Data Linked with Conceptual Component  Groupings/Hierarchies Built On Configuration Component  Lattice, Model Data Associated with Configuration Component  Measurements, Calibration, Traveler, Maintenance Data with Physical Component  Implementation Approaches Set of Tables Set of Properties Core Schema Summary V. Vuppala, Controls DB Meeting, Slide 24

25. Behavioral: Controls V. Vuppala, Controls DB Meeting, Slide 25

26. Behavioral: Model V. Vuppala, Controls DB Meeting, Slide 26

27. Naming V. Vuppala, Controls DB Meeting, Slide 27

28. Traveler V. Vuppala, Controls DB Meeting, Slide 28

29.  Two “Flavors” of IRMIS Systems Controls-based IRMIS (IRMIS-C): Excellent GUI, stable releases Physics-based IRMIS (IRMIS-P): Physics schemas, data loading scripts  IRMIS-P closer to FRIB Schema Configuration-component – install Physical-component – component Conceptual-component – component-type  Ideal situation Everyone agrees to one core schema The GUI can be “ported” to new schema  What all changes are need to IRMIS-C and IRMIS-P? Mapping to IRMIS V. Vuppala, Controls DB Meeting, Slide 29

30.  Possible solutions RDBMS (Relational Database Management System) »Oracle direct path, partitioning »Integrated access SciDB: open source high performance database for scientific problems Hypertable: open source distributed high performance data storage Cassendra: open source distributed high performance database (apache)  Other labs BNL looking into SciDB and Hypertable BNL goal: 1M PVs at 1hz forever SLAC has tried RDBMS SNS is using Oracle (93K PVs) Archiver, Slide 30V. Vuppala, Controls DB Meeting

31. Status, Slide 31V. Vuppala, Controls DB Meeting  IRMIS Mapping  Gaps, and how to bridge them  Refine schemas  Components, lattice, configuration, groupings, security etc  Cables, MPS, Save/Restore etc  Properties and attributes for all classes  Installing MySQL Community Edition  MySQL Cluster to be Tested in Parallel  Generic Design Documentation  To Do: MPS, Interlocks, Operations Modeling

32. 1.Understand and accept that you will make mistakes. 2.You are not your code. 3.No matter how much "karate" you know, someone else will always know more. 4.Don't rewrite code without consultation. 5.Treat people who know less than you with respect, deference, and patience. 6.The only constant in the world is change. 7.The only true authority stems from knowledge, not from position. 8.Fight for what you believe, but gracefully accept defeat. 9.Don't be "the guy in the room." 10.Critique code instead of people—be kind to the coder, not to the code. Ten Commandments of Egoless Programming V. Vuppala, Controls DB Meeting, Slide 32