Large Scale RealTime Embedded Systems

1
Large Scale Real-Time Embedded Systems
National Science Foundation
Collaboration

• University of Illinois
• Vanderbilt University
• Syracuse University
• University of Pittsburgh
• Fermi National Accelerator Laboratory

2
Designing and implementing very large-scale
real-time embedded computer systems

• Ultra high computational performance through use
  of parallel hardware architectures
• Functional integrity via distributed,
  hierarchical monitoring and control
• Highly available, dynamically reconfigurable,
  maintainable, and evolvable architecture

3
Application Trigger and Data Acquisition System
4
The Problem

• Large system
• 100s (1000s) of FPGAs in the muon (pixel)
  trigger
• 250 (2500) DSPs
• 2500 processor Linux farm
• Something will fail !
• How do we get the best/most physics?
• Detection, mitigation, graceful degradation

5
RTES Solution
Modeling
Analysis
Resource
Reconfigure
Performance Diagnosability Reliability
Synthesis
Fault Behavior
Algorithms
Feedback
Design and Analysis
Synthesis
Runtime
Soft Real Time
Hard
6
ARMOR-based Hierarchical Error Management

• Multithreaded processes composed of replaceable
  building blocks (elements)
• Provide error detection and recovery to
  applications
• Hierarchy of ARMORs forms self-checking runtime
  environment

7
Current Testbed
PC - Windows OS
DSP - BIOS
PC - Linux OS
DATA
Physics Application
EPICS Graphical Display System
Gateway
COMMANDS
Very Light Monitor Agent
TCP/IP
TCP/IP
ARMOR Microkernel
ARMOR Microkernel
Msg Parser
DSP Interface
Recovery Policy
Msg Parser
EPICS Interface
Recovery Policy
Local Manager ARMOR
Local Manager ARMOR
Daemon
Daemon