Ethernet is known for asynchronous and non-deterministic communication behavior. With new synchronization capabilities and possibility to send synchronous messages, the application domain of Ethernet can be further extended for embedded systems. In critical networks and embedded systems, the synchronization is particularly important (see comparison IEEE1588 and SAE AS6802), as it becomes the cornerstone of system operation.
The list below shows where synchronization and jitter control (at network level) support system design objectives. One of essential features is the capability to design systems which can be less complex and less costly to design, integrate, verify, maintain, reconfigure and reuse (lower NRE/RE).
In addition, new advanced integrated architectures can be designed to reduce electronic control unit (ECU) cost, minimize ECU number, reduce wiring and weight. The synchronization supports efficient embedded system virtualization, which means that integrated functions can be packed in smaller footprint (less memory, less processing power, less network bandwidth).
The virtualization also means, that many functions can reside on one ECU. For real-time systems, it is possible to virtualize deterministic real-time functions, and the system time (synchronization) with µs-precision is required to make it viable at low hardware costs. Low architecture and hardware costs (RE) are essential for high-volume embedded applications with Ethernet.
Benefits of System Time, Synchronization and Jitter control
- real-time operation in fast processes and controls
- periodic sampling of sensors
- minimized jitter between periods
- control loops with sampling time higher than >30-40Hz, upto Nx10kHz
- minimized SW code size for processing and sampling
- simpler smart sensors
- design of robust redundant systems with simplified redundancy management
- more than 50-60% of code in critical systems (e.g. avionics) is used for maintaining redundancy and recovery
- asynchronous networks require much more complex approaches for recovery and redundancy management
- reducing system complexity
- defining unambiguous system interfaces (interactions!) for different distributed functions
- deterministic communication and coordination of distributed functions
- reduce uncertainity, variability and prevent unintended system states
- minimize transient system faults and NFF (no fault found) in the field
- support for security/safety
- NEW ADVANCED ARCHITECTURES: support for efficent integration of many functions on few computing modules
- efficent use of network bandwidth and computing resources
- sharing and isolation of network bandwidth and computing resources
- efficient virtualization of embedded systems