I did a little searching with the Google and found this excellent and quite readable history of channel coding, with some nice graphs illustrating how close current technology has come to the Shannon bound, see for example Figure 12.
http://arxiv.org/pdf/cs/0611112.pdf I actually took a Coding Theory course from one of the authors, Daniel Costello, when he was a professor at IIT. I am fascinated by coding theory because some of the revolutionary developments like convolutional coding, Viterbi decoding, and turbo codes are quite recent. As the article notes, modern turbo codes come within less than 1 dB of the Shannon bound. From: Ken Hohhof Sent: Thursday, January 15, 2015 11:04 PM To: af@afmug.com Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous If I remember correctly, uncoded QAM is about 8 dB from the Shannon bound. Most licensed radios include some FEC but not ARQ because of the latency hit with ARQ. So some of that 8 dB gap is already made up with coding. There is maybe 3-5 dB left on the table. None of it low hanging fruit. Information theory says with sufficiently complex codes it is theoretically possible to approach (but not exceed) the Shannon bound. (note some assumptions like additive white gaussian noise) What it doesn’t say is how to design these codes, what the computational complexity would be, or significantly, what latency is required. So in the example of NASA deep space probes, latency doesn’t matter. For that matter, neither does decoding complexity, because they can use supercomputers on earth for decoding. So if a block code uses a block size equal to an hour of transmitted data, all of which must be received before you can decode anything, it doesn’t matter. Licensed links however are expected to be very low latency, <<1 msec. You cannot use extremely high complexity FEC, and you can’t use ARQ at all. It really does come down to modulation and channel width, like Daniel’s coffee mug chart. And how many streams you can get per radio head and per antenna. There are some tweaks, but I suspect no huge breakthroughs. Like voiceband modems that peaked with V.92 modulation and V.44 compression and basically that was the end, there was nothing more to be gained, or nothing worthwhile. From: Mike Hammett Sent: Thursday, January 15, 2015 10:07 PM To: af@afmug.com Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous I just actually dug up the numbers and actually, the LTE receive sensitivity that Patrick was talking about are within a dB or two of the SIAE datasheet I have open at 64 QAM. He then talks about 8.5 dB of additional receive gain via 4x4 and HARQ. *shrugs* Every dB helps. ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com -------------------------------------------------------------------------------- From: "Daniel White" <afmu...@gmail.com> To: af@afmug.com Sent: Thursday, January 15, 2015 9:52:27 PM Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous Licensed vendors can’t use many of the tricks unlicensed, or low frequency, vendors get to use. The space is regulated much differently. Daniel White | Managing Director SAF North America LLC Cell: (303) 746-3590 Skype: danieldwhite E-mail: daniel.wh...@saftehnika.com From: Af [mailto:af-boun...@afmug.com] On Behalf Of Mike Hammett Sent: Thursday, January 15, 2015 7:54 AM To: af@afmug.com Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous Well right, we need more bandwidth per link, but also that in many areas there's no links to be had, so we have to make do with what we have. For instance in that spot I'm working on the GigE links for, there's no 6 GHz, 11 GHz and only 2x horizontal 18 GHz licenses available. It's not exactly suburbia or HFT alley, either. I think the next step will have to be IP20 type solutions from more vendors. Patrick is talking all kinds of engineering finesse that lowers how much signal LTE needs to work well by like 10 - 15 dB. I wonder how much of that would help licensed links. The thought there is that having more signal to work with means that higher QAMs aren't as fragile as without that fanciness. Roadmap to 10G via licensed, anyone? ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com -------------------------------------------------------------------------------- From: "Ken Hohhof" <af...@kwisp.com> To: af@afmug.com Sent: Thursday, January 15, 2015 8:41:42 AM Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous All good points, not sure they address Erich’s issue of needing more bandwidth per link without multiple antennas per link on the tower. I think he’s saying an IP20 class radio is too expensive, and over the next few years lots of us are going to need that kind of radio. From: Mike Hammett Sent: Thursday, January 15, 2015 8:27 AM To: af@afmug.com Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous Pull all of the active licenses with old ass gear and feed their contact information to the vendors? ;-) Maybe newer gear for hte existing guys would cut down on how much they need. ;-) ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com -------------------------------------------------------------------------------- From: "Mike Hammett" <af...@ics-il.net> To: af@afmug.com Sent: Thursday, January 15, 2015 8:25:39 AM Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous 10 GHz unlicensed would help. I sometimes think the license protections are a bit excessive. Maybe loosen the geographic restrictions on 7 GHz and 13 GHz? Maybe some effort into getting 4 GHz and other legacy PtP bands opened back up? They don't have a ton of room and don't allow for huge channels, but some is better than none and maybe since the HFT guys care more about lower latency and less about throughput, they''ll build longer hops and leave our 11 and 6 gig alone? ;-) ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com -------------------------------------------------------------------------------- From: "Mike Hammett" <af...@ics-il.net> To: af@afmug.com Sent: Thursday, January 15, 2015 8:18:58 AM Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous You could drive around to all of the HFT sites and look for links that are no longer there, but still licensed. Document. Come back 30 days later and document again, submit to the FCC (or wherever Liz says is best) and have those licenses revoked. ;-) But yeah, I do like Ceragon's 4x4 setup. Two radio heads, four transmitters, two licenses, tons of bandwidth. ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com -------------------------------------------------------------------------------- From: "Erich Kaiser" <er...@northcentraltower.com> To: af@afmug.com Sent: Thursday, January 15, 2015 8:15:57 AM Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous The conversation goes deeper. In several areas we are out of spectrum, it may be due to HFT or just plain out. So most WISPs if they have deployed Licensed, they have to deal with what spectrum they have. The question is, where do you see the licensed backhaul market in the next few years, are they just going to be adding qams or finding other creative ways to add capacity? I am not in the WISP business anymore so I have decided to try and focus on the things that drove me nuts to help others. Erich Kaiser North Central Tower Consulting er...@northcentraltower.com Office: 630-621-4804 Cell: 630-777-9291 On Thu, Jan 15, 2015 at 7:33 AM, Daniel White <afmu...@gmail.com> wrote: Your issue is really then with promo’s and capacity keys – which is marketing and different ways to make revenue on a product. Just like the PtMP space, not all PtP vendors do capacity keys and play those games either. Daniel White (303) 746-3590 From: Af [mailto:af-boun...@afmug.com] On Behalf Of Erich Kaiser Sent: Wednesday, January 14, 2015 9:47 PM To: af@afmug.com Subject: Re: [AFMUG] Licensed backhaul pricing - still ridiculous From my experience the price has not changed very much. Someone needs to take the reigns on the market. Even with certain companies throwing you a promo price, if you really look at it fully loaded, the price is still high for what you get. On Wed, Jan 14, 2015 at 10:35 PM, Jason McKemie <j.mcke...@veloxinetbroadband.com> wrote: I think we're already seeing it. They're not wifi chipset based radio kind of prices, but they can be found for less than half of what you could get them for just a few years ago. On Wed, Jan 14, 2015 at 10:30 PM, Erich Kaiser <er...@northcentraltower.com> wrote: After several years, when will we see Licensed radios come down in price? There is so much margin in these things. Its ridiculous....