VXS Processor Mesh: Bringing VXS Performance to the Next Level

One common misconception about VXS is that can utilize only Star and Dual Star topologies. In other words, have only one or two switch cards running the fabric and the traces going to payload cards. One of Elma Bustronic’s earliest VXS design ideas was to incorporate a “switchless mesh” or “payload mesh”, where a point-to-point mesh interconnection ran across three VXS payload slots. Therefore, the mesh ran across the P0 section of the three slots in the backplane. This configuration is ideal for prototyping and testing VXS, is it does not require the use of switch cards to run the fabric.

However, it is possible to have the kind of mesh connectivity that runs across the connector configurations found in the switch card slots. This provides much more pins for either mesh or redundant mesh configurations or for more I/O. Processor Mesh is a new VXS innovation created by Elma Bustronic in conjunction with Tek Micro and Qinetiq. The concept is quite similar to what a mesh section in 6U VPX (VITA 46) would look like. VXS Processor Mesh offers a defined pinout for the mesh section of the backplane. The initial backplane Elma Bustronic developed offers a combination of legacy VME64x slots, VXS payload slots, a VXS switch slots, and VXS Processor Mesh slots. The result is a significant jump in bandwidth. While VXS in the star topologies offers up to 20-30 Gbps, VXS Processor Mesh is estimated at 112 Gbps.

Elma Bustronic and our partners proposed the Processor Mesh concept to VITA and the VITA 41.7 subcommittee for Processor Mesh has been formed. Early pinouts of the initial Processor Mesh backplane are currently available. An illustration shows the topologies in more detail.