A mesh wireless network node must perform three functions:
- serve client devices,
- receive traffic from another mesh node (mesh ingress) and
- transmit traffic to another mesh node (mesh egress).
Single-radio nodes (or even dual-radio nodes, in which one radio is used for client devices and the other for mesh backhaul) pose performance problems and hinder scalability.
Multi-radio mesh technology solves this problem for large-scale wireless deployments, particularly those in which real-time applications require roaming voice, video and data. Radios are dedicated for each function in the mesh - backhaul ingress, backhaul egress and client access. This provides dedicated mesh links for backhaul traffic and client coverage. 802.11a is generally used for backhaul traffic, and 802.11b/g for client coverage.
A single radio makes a bottleneck
When a single-radio mesh is used for both incoming and outgoing traffic, throughput is halved because a radio cannot transmit and receive simultaneously, and must swap roles. Another problem is that every mesh link must be on the same radio channel. That means when one radio is transmitting, its neighbors must all be in listening mode. This problem is amplified across the mesh, and after a few hops the architecture is slowed to the point where it no longer efficiently supports voice or data.
However, a wireless multi-radio, or structured, mesh approach offers several dedicated-link interfaces and at least three radios per network node. Because each radio performs only one function, there is no role swapping, eliminating throughput degradation. Multi-radio mesh networks allow for dedicated backhaul links that can transmit and receive simultaneously because each link is on a separate channel.
Why have six radios?
A typical mesh configuration might have six or more radios in a single node that can be allocated between serving client devices, mesh ingress and mesh egress. This six-radio wireless network node can support up to three user-access radio modules in the 2.4-GHz or 5-GHz band and up to three backhaul radio modules in the 2.4-GHz or 5-GHz band.
The ingress radio must be designed to receive associations from wireless clients (user access) or other mesh nodes (egress) in separate radios. The egress radio must create a link to another mesh node for relaying backhaul traffic, based on the best possible path to the wired network, or broadband termination point. Many factors can be optimised across the mesh links with a multi-radio approach, including channel optimisation, round-trip delay, signal strength and packet routing.
Dedicated radios make for structure
By dedicating radios to each mesh function, structured multi-radio mesh networks increase throughput and reduce latency, enabling the networks to accommodate voice applications and fast-roaming handoff, as well as video and other real-time applications. Multi-radio mesh is designed around Layer 2 switching, further reducing latency and overhead while improving multi-hop performance.
A structured multi-radio wireless mesh network also allows for the use of multiple, separate sectorised antennas that send signals in different directions, each on different channels, and all at the same time. This cellular-like coverage enables simultaneous, collision-free transmission among clients associated with this type of architecture. As a result, more users can associate with the same node at longer ranges and attain a higher overall throughput, because there is less contention with other users.
Multi-radio Wi-Fi mesh networks solve the inherent Wi-Fi multi-hop throughput and latency dilemma caused by the half-duplex nature of 802.11. The performance gains achieved are ideal for supporting real-time voice, video and data over multiple hops.
Irani is vice president of marketing and strategy for Strix Systems. This article appeared first in Network World.