World's First Fully Programmable and Open-Source Massive-MIMO Platform

Re:World's First Fully Programmable and Open-Source Massive-... by LockAndLol on Sunday August 02, 2020 @09:23AM (#24028542)

Does this mean a real possibility for user-defined, community 5G networks? If so, that could usher in a new age of mesh networks to democratize the long-range wireless network space. If the hardware is cheap and a protocol were used that's anonymous and encrypted by default, this could be a game changer.

From an idealistic newbie's perspective, I imagine the possibility of users buying routers with this tech, plugging them into their fiber or copper connection, and creating or expanding an existing 5G meshnet + a home net. Maybe it could drastically drop costs for 5G network operators as they wouldn't need to put up extra receivers anywhere. They could simply offload that burden to the consumer. Most routers are lent out anyway, so not only would they make money, consumers wouldn't need to worry about which operator they chose because they'd all be in the same meshnet anyway.

But eh... I'm sure there's some big thing I'm missing besides human greed.

[ Parent | on HN ]

Re:World's First Fully Programmable and Open-Source Massive-... by namibj on Sunday August 02, 2020 @11:23AM (#24029250)

Well, a LimeSDR based E-UTRA (LTE) Band 38 2x2 eNodeB could be had for a couple hundred USD, depending on how cheap you can get the Host computer and clock source.

I'd consider just injecting the clock via the fiber (running a dedicated single-mode fiber from the next PoP (up to ~10 km away) to the eNodeB, potentially exposing Ethernet service from the base station, just like the typical monolithic DOCSIS wifi router), via one of two ways:

(1) run a separate wavelength via CWDM and just AM-modulate a sinusoidal reference oscillator output onto it, or

(2) anchor the clock for the data link from the PoP to the eNodeB to your reference oscillator, and just tap your reference clock (to feed into the local VCXO-based master-PLL) off of the data link receiver's clock recovery circuity. Some mild phase noise is acceptable thanks to the VCXO.

[ Parent | on HN ]

Re:World's First Fully Programmable and Open-Source Massive-... by apcragg on Sunday August 02, 2020 @04:01PM (#24031240)

Well, for one you're not going to get a very wide coverage range or throughput with the ~1 watt of power you're limited to in the ISM bands. Your router will also have to accept interference from 802.11 routers, Bluetooth, etc devices. There is a lot of on going research into dynamic spectrum allocation and cooperative scheduling which this platform is actually supporting but that still requires federation which your use case seems to be against.

[ Parent | on HN ]

Re:World's First Fully Programmable and Open-Source Massive-... by crote on Sunday August 02, 2020 @10:08AM (#24028798)

That sounds an awful lot like what Fon Wireless is doing.

[ Parent | on HN ]