We’ve heard of how the cellular industry plans to invade unlicensed spectrum by adapting LTE for the purpose. Mimosa is a Wi-Fi vendor choosing to modify that technology for best effect in both unlicensed and licensed bands. Jaime Fink, their CTO, explains their chosen market segment and why he thinks their technical approach will win out financially in the long run.
Targeting the Yellow Zone
Mimosa is a US based wireless equipment vendor offering a “fibre replacement” service for suburban areas. Jaime believes there are many towns of reasonable size (say 800 homes per square km) which don’t fit the DOCSIS cable, VDSL or Fibre business models. They may have ADSL broadband service but have few other choices for faster speeds. He believes these “Yellow Zones” can be served cost effectively using unlicensed spectrum and comprise up to 20% of the addressable broadband Internet market. He sees huge potential in both developed and developing countries.
The easiest transition points are to offer either:
- a pureplay fixed broadband service that directly competes with wireline and cable providers.
- a public Wi-Fi service, especially in developing markets with little wireline infrastructure
He’s developing technology that can deliver sustained 100Mbps datarates (burst rates up to 500Mbps) to each home at a capital cost of $150 per home served. This could be offered exclusively to residents or resold as public Wi-Fi for $10-$20/month, possibly with a slower and/or time-limited basic free service.
Distribution would be directly between rooftops of homes, with those hosting an aggregation hub given discounted or free broadband service in compensation. A topology with as few as 10 homes per hub would allow useful speeds of up to 100Mbps per user.
Adapting mass market Wi-Fi products
While the customer connects using standard Wi-Fi, Mimosa have adapted the Wi-Fi protocols to increase efficiency and create a low cost backhaul network.
The "Mimo" in Mimosa relates to Multi-User Multiple Input Multiple Output, allowing a single access point to share/reuse the same spectrum in multiple directions. It also helps to overcome obstructions and reflections in the signal path.
Where LAA is an adaptation of standard LTE to play nicely with other Wi-Fi, Mimosa’s proprietary solution adds features to Wi-Fi to give it some of the capabilities seen with LTE. I suspect this breaks aspects of the “Listen Before Talk” features of Wi-Fi and may not be accepted by regulators in some countries.
Unlike 802.11ac which spreads over up to 160 MHz of spectrum, their system squeezes the most out of a single 20MHz channel. They synchronise their access points across the network and recycle the spectrum. Multi-User MIMO is used extensively with plans for 8x8. The uplink is constrained to Single User MIMO until the next generation of 802.11ax.
By basing their solution on commodity Wi-Fi chipsets (from Quantenna), Jaime believes this give them a significant cost advantage over LTE small cells which ship in relatively smaller volumes. Quantenna have gained substantial market share in residential Wi-Fi access points but not yet so dominant in Enterprise or Carrier grade products.
Using 5GHz for long distance backhaul
Mimosa has also been shipping kit for use in point-to-point long distance backhaul for Wireless ISPs.
Jaime thinks that much of this WISP backhaul traffic will move out of the unlicenced 5GHz spectrum which is becoming more congested. Frequencies at 24GHz and 10/11GHz are good alternatives where available.
The 5GHz band also suffers from heavy loss through trees. One way around this issue is for Wireless ISPs to put in more rooftop sites and route the traffic in different ways. Mimosa have created a simulation planning tool to help identify which customer sites would best become an aggregation point. A cluster of 10 homes in line of sight can be cost effective, delivering fibre like 100Mbps rather than todays typical 3Mbps service to each.
It’s not just the cellular equipment vendors who are proposing unusual and currently non-standard ways of using unlicenced spectrum more efficiently. Mimosa has developed a proprietary wireless backhaul solution that complements standard Wi-Fi access points to deliver low cost broadband service.
In the context of the Wi-Fi vs LTE-U/LAA debate, it’s interesting to see an alternative approach. Mimosa argue that their design promises significant spectrum efficiency through heavy reuse.
Basing their design using mass market Wi-Fi using commodity chipsets reduces the component cost. A concern I have is that their technology modifications may not play nicely with Wi-Fi and be restricted by regulators when used for unlicensed bands in some countries.
Their primary market is to compete with wireline broadband service, especially in those areas limited to ADSL over long copper wires. At this stage they are not offering a mobile or cellular service, and their primary customers are Wireless ISPs.
The growing congestion of the unlicensed bands (including at 5GHz) means that much of the backhaul will be migrated to other licenced spectrum in the long term.