Does the Number of Spatial Streams in 802.11ax Really Matter?

802.11ax has arrived, and with that, you will see many vendors announcing their 802.11ax access points in the coming months. We have discussed numerous capabilities of 802.11ax on this blog, however, there is one question that has come up recently. What is the difference between a 4×4:4 AP and an 8×8:8 AP? Which do I actually need?

Having an increased number of spatial streams hold 2 potential benefits:

1. Achieving higher data rates while communicating with a single client.
2. Achieving higher aggregate performance in a MU-MIMO environment when simultaneously communicating with multiple clients.

However, these two MU-MIMO benefits currently are only attained in testing environments and not in production networks. It has been over 3 years since the first 802.11ac wave 2 access points launched and still hardly any MU-MIMO capable clients exist in the current marketplace and the technology is rarely used in the enterprise. Most current clients do not even support downlink MU-MIMO. Additionally, word on the street is that uplink MU-MIMO will not be supported when the first generation of 802.11ax is certified by the Wi-Fi Alliance.

The most powerful client devices on the market are 3×3:3 (MacBooks, Dell XPS laptops, etc). No device out there benefits from 4 spatial streams, let alone 8 – and it’s very unlikely we are going to see anything more than 4×4 clients in the future. The vast majority of client devices use 2×2:2 MIMO radios.

MU-MIMO requires a number of factors to be ‘in-play’ to operate effectively— such as spatial diversity, where physical distance between the clients is necessary. Most modern-day enterprise deployments of Wi-Fi involve a high density of users:

• Because MU-MIMO requires spatial diversity, a sizable distance between MU-MIMO clients is required to prevent interference. Once again, most modern-day enterprise deployments of Wi-Fi involve a high density of users that is not conducive for MU-MIMO conditions.
• Because MU-MIMO requires spatial diversity, a sizable distance between the MU-MIMO clients and the AP is also necessary.

MU-MIMO requires transmit beamforming (TXBF) which requires sounding frames. The sounding frames add excessive overhead, especially when the bulk of data frames are small. MU-MIMO would only be a favorable option in very low density, high bandwidth environments, with a small number of users requiring extremely high throughput.

*Keep in mind that 8×8:8 APs will initially be more expensive and will always require more power. In fact, we are starting to see power requirements for 8×8:8 APs that exceed 802.3at power capabilities. Even switches that support PoE+ will not be able to provide adequate power and therefore a special power injector will be needed.

Most industry experts believe that multi-user OFDMA will be the most relevant technology that 802.11ax offers. Regardless of stream count, all APs will support the same number of 802.11ax OFDMA clients. OFDMA technology makes better use of the available frequency space by subdividing the channels into resource units for simultaneous multi-user transmissions both downlink and uplink. Even with a 20 MHz channel, we can talk simultaneously to up to 9 client devices and theoretically up to 37 with an 80 MHz channel (although enterprise customers should rarely use anything above 20 MHz).

The result is an effective multi-client performance. With dual 5 GHz capability, vendors can effectively double 5 GHz capacity and improve the effectiveness of a legacy environment (that does not yet have 802.11ax devices) while simultaneously doubling multi-client performance potential via OFDMA.