The concept behind an Ethernet Fabric is a merge of a few different ideas and technologies. The term “Silicon Fabric” refers to the chip that performs high speed forwarding of data from input to output.
In a storage network or SAN, each FibreChannel switch had a high performance silicon fabric, but also each and every switch was part of a whole. Since the host must register with the switch, and that switch tells each other switch about the host, the entire SAN was perceived, and marketed, as single fabric.
The art of the converging storage and data onto a single network relies on detecting different traffic types and then forwarding them differently, according to their needs. Storage traffic wants to be lossless, reliable and low latency. Data traffic needs bandwidth and flexibility. Ethernet switches have similar silicon fabrics, but are now performance boosted to the same level as SAN hardware.
The DCBX standards for Ethernet mean that this new class of switches signal to each other using Enhanced Transmission Selection (ETS IEEE802.1Qaz) to map traffic into traffic shaping groups using Priority-based Flow Control (PFC IEEE802.1Qbb). PFC also signals between switches when congestion occurs.
Previously, Ethernet switches were autonomous systems sending only STP message. Each switches received STP message and built it’s own forwarding configuration. Now, each Ethernet switch is part of a system, and signals its neighbour as part of unified whole. Like a SAN Fabric, or a piece of cloth.
So now we call Data Centre networks “Ethernet Fabrics” as standard terminology.
Other posts in the series
- ◎ What's Happening Inside an Ethernet Switch ? ( Or Network Switches for Virtualization People )
- Tech Notes: Juniper QFabric - A Perspective on Scaling Up
- Switch Fabrics: Input and Output Queues and Buffers for a Switch Fabric
- Switch Fabrics: Fabric Arbitration and Buffers
- What is an Ethernet Fabric ? (This post)
- What is the Definition of a Switch Fabric ?
- Juniper QFabric - My Speculations