GE Fanuc VMIPCI-5565 Ultra High Speed Fiber Optic Reflective Memory

Functional Features

- High-speed, easy-to-use fiber optic network (serial rate of 2.12 Gbaud)

- Data written to the memory of one node is also written to the memory of all nodes on the network

- Up to 256 nodes

- Connections up to 300 meters to multimode fiber and up to 10 kilometers to single-mode fiber

- Dynamic packet size, 4 to 64 bytes of data

- Transfer rates 47.1MB/sec (4-byte packets) to 174MB/sec (64-byte packets)

- 64MB or 128MB SDRAM Reflective memory

- Two independent DMA channels

- Any node on the network can generate an interrupt in any other node on the network or in all network nodes with a single command

- Error Detection

- Redundant transfer mode for additional error suppression

- No processor overhead

- Processor is not involved in network operation

- PCI 64-bit 66MHz Transport

- Compliant with PCI Revision 2.2

- Includes VMISFT-RFM2g network and shared memory drivers.

Features

Description: The VMIPCI-5565 is the PCI member of the VMIxxx-5565 family of reflective memory (RFM) real-time networking products from GE Fanuc Embedded Systems.

Other members of this family include the VME-compatible board VMIVME-5565 and the PCI mezzanine card (PMC) VMIPMC-5565.

All three products are network compatible and can be integrated into a network in any combination.

This family of products allows computers, workstations, programmable logic controllers (PLCs) and other embedded controllers with different operating systems or no operating system at all to share data in real time.

Reflective Memory boards are like shared memory for local nodes.

Any level of software, including the application program itself, can write to or read from the memory.

Data written to one node's Reflective Memory is transferred by the network hardware to all other nodes and placed in the same addresses on those nodes' Reflective Memory boards.

This data transfer does not require the involvement of a processor on any node. With this system, all nodes on the network have a local copy of the shared data that is immediately accessible.

Product Overview: The Reflective Memory concept provides a very fast and efficient way to share data between distributed computer systems.

The VMIPCI-5565 Reflective Memory Interface allows data to be shared between up to 256 independent systems (nodes) at rates up to 174MB/s. Each Reflective Memory board can be configured with 64MB of memory.

Each Reflected Memory board can be configured with either 64MB or 128MB of on-board SDRAM. local SDRAM provides fast read access times to stored data.

Write data is stored in the local SDRAM and broadcast to other Reflective Memory nodes via a high-speed fiber optic data path.

Data transfers between nodes are software transparent, so there is no I/O overhead.

Transmit and receive FIFOs buffer data during peak data rates to optimize host and bus performance and maintain high data throughput.

Reflective memory also allows interrupts to one or more nodes by writing to byte registers.

These interrupts (four levels, each user-definable) can be signaled to synchronize system processes or to follow any data.

The interrupts always follow the data to ensure that the data is received before the interrupt is acknowledged.

Each node in the system has a unique identification number between 0 and 255.

The node number is determined by placing jumpers on the board during hardware system integration.

Software can read the node number by accessing on-board registers. In some applications, this node number helps determine the function of the node.

Link Arbitration: The VMIPCI-5565 system is a fiber optic daisy chain ring as shown in Figure 2.

Each transmission travels from one node to another until it wraps around the ring and reaches the originating node.

Each node retransmits all transmissions it receives except those it originates.

A node can insert transmissions between passed transmissions.