Enhanced Ring Fault Tolerance

Overview

This section describes the Enhanced Ring Fault Tolerance (ERFT) and Single Ring Redundancy (SRR) features.

The ERFT feature enhances the EXNET ring robustness, performance, and tolerance to unexpected faults.

The EXNET-ONE card has the ability to signal to its adjacent nodes when it has opened or looped back its port. This one feature enables the ring controller software to support online recovery from a single point of failure, seamless single ring redundancy switchover, and improved performance.

Primarily, ERFT supports the ring interface controller (EXNET-ONE card) component, which is responsible for establishing and maintaining the EXNET ring using the ERFT ring mode.

The ERFT ring mode supports the EXNET-ONE ring interface controller card, activating full ERFT support. This includes enhanced fault detection, isolation, and recovery, increased performance and response time, enhanced SRR with seamless switchover support, and enhanced live node insertion (passive addition).

Features

The following enhanced fault tolerance features are provided by the EXNET ring controller software:

EXNET Live Node Insertion (Passive Addition)

The EXNET live node insertion (passive addition) has been merged into the ERFT ring mode state machine. The EXNET-ONE card synchronizes the two adjacent nodes to loop back and expand their ports, so as to not bring down the ring. The EXNET-ONE card also has the ability to turn off the fiber optic port light to synchronize and signal between adjacent nodes. Passive addition, using the ERFT state machine, allows physical, link, and data validation on each port independently.

Single Ring Redundancy (SRR) with Seamless Switchover

While remaining transparent to the host, SRR performs a switchover seamlessly, without dropping calls due to the fact that a node can removed itself (or be removed) from the ring, allowing for a standby node to take over without disrupting the ring or local bus.

Debugging Capabilities

A real-time logging function has been created to trace various conditions. Also, since the logging function requires no resources and a small amount of CPU overhead, ISRs are permitted to log events. The logging function is similar to "printf", except that a log type is also supplied for filtering. Each event is time stamped with a resolution of five milliseconds and stored in a real-time circular buffer that is retained through a push button reset. While the ring is up, each node can synchronize their times to analyze timing sensitive scenarios.

The ring monitor is accessible via the debug console. At any time, monitor statistics can be queried, including TX and RX slips, node accessibility, and acquired failures.

ERFT Ring Mode

After the EXNET-ONE card has been assigned to a Logical Ring ID, the host uses the EXNET Ring Configure message to enable the ERFT Ring Mode, which supports the following functionality:

• Multi-ring Redundancy/Expanded Switching

• Ring Timing/Mastership Configurability

• Enhanced Ring Initialization

• Enhanced Live Node Insertion/Removal

• Runtime Software Link Validation

• Enhanced Single Ring Redundancy ("Hot" Standby)

• Improved Performance

• Improved Ring Status Report Generation

• Improved Mastership Re-arbitration Logic