Why Broadcom would pass on Ruckus

Shame that Broadcom is passing on Ruckus. That would have been an interesting set of conversations.

Broadcom is buying Brocade, although it will not be keeping the IP networking business, which includes Ruckus, maker of WiFi access points and controllers.

Although the main reason is clearly strategic – why would Broadcom want to be vertically integrated with a company that competes with many of its biggest customers in WiFi  – there is another side-angle. Ruckus has not always been entirely complementary about Broadcom’s chip designs and the products its customers build upon them.

Note that Ruckus uses Qualcomm Atheros chips as its silicon platform. One proposed reason for this is because it can customise the Atheros designs, for example to hook its proprietary BeamFlex technology into the Atheros chips.

In fact, in recent times Ruckus has made it pretty clear it does not see a purely standards-based beam forming approach (known as TxBF) in 802.11ac to be the best approach. That is why it wrote this article in 2013 stating that those who think they are getting standards based beam forming from Broadcom may not be getting what they think they are getting.

“Beware that not all 802.11ac access points will be created equal.  Many using the Broadcom chip will suffer from some limitations such as the inability to support more than 50 encrypted clients. And without an onboard CPU in the Broadcom .ac, all Wi-Fi functions are processed by the host’s (AP) CPU. This is far less efficient because the offloading prevents the AP CPU from using low-power states.”

In 2014 the company was again blogging about how standards-based beam forming behaves better in combination with Ruckus’ own BeamFlex technology.

“Because it is a radio technique, effective TxBF DOES require client support (something a lot of people fail to understand). Consequently, 802.11a/b/g/n clients miss out on the perks. And some 11ac clients do not support TxBF. Looking at the pervasive adoption of 11n, we should not expect all (or even most) clients to support TxBF even by the end of 2016. TxBF must also tradeoff with spatial multiplexing. The same transmitters cannot be used for both. In order to be effective, TxBF systems should have double the number of transmit antennas as spatial streams.”

And only just over a month ago Ruckus was again reiterating the point that beam forming and BeamFlex are a “match made in heaven”.

“A popular misconception, spread by our competitors, is that TxBF accomplishes the same end goals as BeamFlex, rendering BeamFlex useless. They are wrong. While the chip-level beamforming is beneficial, it does not provide the same benefit set as Ruckus BeamFlex.”

Although these articles don’t say so outright, vendors basing their products on Broadcom chips are amongst the targets.

Broadcom has said that it wants to move the Brocade IP Network business on: it would be nice to think that the Broadcom WiFi chip team and the Ruckus developers will get at least some time together to share notes on all the above.

In the mean time, who could be in the market for a business formed of Brocade’s “New IP”softwarised Router/Switch businesses, in combo with the Ruckus WiFi business. Could they even be split up? If so, could we see Juniper Networks once more touted as a possible landing place for Ruckus – giving it a mirror access point capability to Cisco.