D. A. Dohan
Abstract:The Advanced Photon Source is now in its fifth year of routine beam production. The EPICS-based [1] control system has entered the phase in its life cycle where new control algorithms must be implemented under increasingly stringent operational and reliability requirements. The sheer volume of the control system (~270,000 records, ~145 VME-based input-output controllers (IOCs), and ~7,000,000 lines of EPICS ASCII configuration code), presents a daunting challenge for code maintenance.
The present work describes a relational database that provides an integrated view of the interacting components of the entire APS control system, including the IOC low-level logic, the physical wiring documentation, and high-level client applications. The database is extracted (booted) from the same operational CVS repository as that used to load the active IOCs. It provides site-wide decision support facilities to inspect and trace control flow and to identify client (e.g., user interface) programs involved at any selected point in the front-end logic. The relational database forms a basis for generalized documentation of global control logic and its connection with both the physical I/O and with external high-level applications.
What problem does this paper attempt to address?
The problems that this paper attempts to solve are: in the Advanced Photon Source (APS) control system, in the face of increasingly stringent operation and reliability requirements, how to effectively manage and maintain a huge control software code base (about 270,000 records, 145 VME - based input - output controllers (IOCs), and about 7,000,000 lines of EPICS ASCII configuration code). Specifically, the paper focuses on the following aspects:
1. **Challenges in code maintenance**:
- The large scale of the system and the high complexity of the code lead to difficulties in maintenance and update.
- New control algorithms need to be implemented without disrupting existing functions.
2. **Requirements for decision support**:
- An integrated view is required to understand and track the interaction relationships among various components in the entire APS control system, including low - level logic, physical wiring documents, and high - level client applications.
- Provide a general documentation method for global control logic and its connection with physical I/O and external high - level applications.
3. **Reliability and performance improvement**:
- Ensure the reliable operation of the system while reducing beam downtime (MTTR) and decreasing the beam loss frequency (MTBF).
- Conduct a detailed analysis of the installed operation control system to understand software behavior and its impact on each component of the control system.
To solve these problems, the paper proposes a decision - support facility based on a relational database, aiming to help application developers of the control system analyze the existing operation control software. By modeling EPICS control software into a relational database, this facility can provide the following functions:
- **Integrated view**: Capture and query the knowledge base of the control system through a relational database, providing a comprehensive view of the entire control system.
- **User query tool**: Develop an application named `pv_search` that allows users to query and browse EPICS records and their field details through a web browser.
- **Data mining**: Utilize the powerful query capabilities of SQL for data mining, such as identifying orphan records, multiply - defined hardware output records, etc.
- **Document integration**: Expand the wiring database to cover all APS control hardware and fully integrate it with the PV database to facilitate cross - hardware and software queries.
These measures are helpful to improve the maintainability, reliability, and performance of the system, thereby better supporting the daily operation and future development needs of APS.