SCRAMNet+ SC150/SC150e Boards

The SCRAMNet+ SC150e cards are backwards compatible with SCRAMNet+ and the original SCRAMNet Classic product except for the GOLD RING communication protocol.

The SCRAMNet+ SC150e products use an enhanced architecture where shared memory read operations bypass the SCRAMNet ASIC for improved performance. Although both PIIO and DMA operations are improved, DMA operations will give the greatest increase in performance. For PIO operations, the overall system performance is also improved as read latency is reduced, the number of PCI retries is decreased, and the PCI bus utilization is improved. Hoever, the performance increase is limited by PIO accesses to the host.

The SCRAMNet+ SC150e board requires one a single slot in the computer chassis.

• A ring topology with 150 Mb/s line transmission rate.
• A "Data-Filter" that allows only data stored in shared memory that has changed to be passed to the network for communications to the other nodes.
• BURST Mode protocol (Error Correction Disabled) with fixed-length message packets of 82-bits.
• BURST PLUS Mode communication based on variable packet size to a maximum of either 256 bytes or 1024 bytes.
• PLATINUM Mode protocol (error correction enabled) with fixed-length message packets of 82-bits.
• PLATINUM PLUS Mode communication based on variable-length message packet size to a maximum of either 256 bytes or 1024 bytes.
• 256-node capacity on each ring.
• No operating or system software required to support network protocol.
• No network-dependent application software required.

Every computer on the network has a constantly updated local copy of all global data, which is passed to all the network computers. The network protocol is implemented in the SCRAMNet+ hardware and therefore no software overhead is required to retrieve this information from the network.

In its simplest form, the SCRAMNet+ SC150/SC150e Network system is designed to appear as general-purpose memory. The use of this memory depends only on the conventions and limitations imposed by the specific host computer system and operating system. On most processors, this means that the application program can use this memory in basically the same way as any other data-storage area of memory. The memory cannot be used as instruction space.

The major difference between SCRAMNet+ SC150/SC150e memory and system memory is that any data written into SCRAMNet+ SC150/SC150e memory is automatically sent to the same SCRAMNet+ SC150/SC150e memory location in all nodes on the network. This is why it is also referred to as replicated shared memory. A good analogy is the COMMON AREA used by the FORTRAN programming language. Where the COMMON AREA makes variables available to subroutines of a program, SCRAMNet+ makes variables available to processors of a network.

When a host computer writes to the shared memory, the SCRAMNet+ SC150/SC150e node host adapter supplies the proper handshaking logic. The shared memory behaves somewhat like resident or local memory.

A software driver is usually not required except for interrupt handling.