Re: [AFMUG] B11

[John Blake]

It's obviously pricier than the B11, but if you want true symmetrical and a
ton of throughput, you could look at the Huawei RTN-380 radios.  These will
do 4Gbps (2Gbps symmetrical) at full licensing, or there are 1, 2, 3Gbps
licensing options.  They use 71-76 GHz and 81-86 GHz and so are super easy
and cheap to get (lightly) licensed. These are very popular with carriers
outside the US, but not so much in the US because of all the
Huawei/Ciscolobby disputes, but that has largely been resolved and they
have been getting traction here.  Let me know if you want more info, we are
Huawei VAR.

John


On Wed, Aug 3, 2016 at 9:54 PM,  wrote:

> Gino,
>
> That seems to be the only configuration that meets the requirement.  The
> way I see it, in a traditional FDD system you would license an XPIC pair of
> frequencies, say 11075 H/V at point A and 11565 H/V at point B.  Assuming
> an 820 running 256QAM you might get 500mbps per polarity per direction, for
> a two-way aggregate of 2gpbs.  You would have licensed 160MHz at each end
> of the system, for a system efficiency of 2gbps/320MHz = 6 b/s/hz.
> If you operate a B11 on the same pair of frequencies, you have to use
> their FD mode which a typical PCN claims will deliver 736mbps.  It would
> seem that this can be viewed as a two-way aggregate rate, because the
> radios still take turns transmitting as in a true TDMA system.  So, the
> system efficiency here is 736mbps/320MHz = 2.3 b/s/hz -- below the
> requirement of 3 b/s/hz.
>
> If you switch the B11 to the normal TDMA mode you will need to transmit on
> the same frequency from each end of the link.  So, in addition to licensing
> 11075 H and V transmitting from point A, you also need to license 11075 H
> and V transmitting from point B, which adds another 160MHz at each end. The
> PCNs show this configuration giving 1.47gbps (again assumed to be an
> aggregate figure due to the TDMA mode).  This is an efficiency of
> 1.47gbps/640MHz = 2.3 b/s/hz again.  However, because you have licensed two
> frequency pairs, each site can also transmit and receive on the unused
> 11565 H and V frequencies.  If you do this then you get 2*1.47gbps/640MHz =
> 4.6 b/s/hz.  This seems to be the only valid configuration, but does this
> take two radios at each end, or just one?
>
> Mike Black
>
> Black & Associates
>
> 727-773-9016
>
>
>  Original Message 
> Subject: Re: [AFMUG] B11
> From: "Gino Villarini" 
> Date: Wed, August 3, 2016 3:38 pm
> To: "Animal Farm" 
> --
>
>
> > but the radio tx in both channels in the both polarites in both ends ...
> >
> > On Wed, Aug 3, 2016 at 1:31 PM, Eric Kuhnke 
> wrote:
> >
> >> No, it is just as spectrally efficient as any 256QAM radio... One 80 MHz
> >> "low" channel in both polarities and one 80 MHz "high" channel in both
> >> polarities, as a typical FDD band plan such as you would use with a
> >> configuration with 2 dishes, 2 orthomode transducers and 4 radio heads
> >> (each radio operating in a single polarity) in a 2+0 configuration.
> >>
> >> You're arriving at the figure of 320 MHz by counting everything twice.
> >>
> >> On Wed, Aug 3, 2016 at 10:28 AM,  wrote:
> >>
> >>> We haven't had the pleasure of coordinating a B11 link yet, but they
> >>> certainly seem to be popular based upon the number of PCNs we are
> >>> seeing. So, a dumb question hopefully based upon a simple
> misunderstanding
> >>> of the numbers: I don't see how this meets the FCC minimum efficiency
> >>> standards for wide channels at 11GHz:
> >>>
> >>> 80MHz channels x H and V = 160MHz, but with high/low pairing you double
> >>> this? to 320MHz occupied per end? So, to meet the 3 b/s/hz requirement
> at
> >>> 11GHz your symmetrical throughput would need to be >= 960mbps. What am
> I
> >>> missing?
> >>>
> >>> Mike Black
> >>>
> >>> Black & Associates
> >>>
> >>> 727-773-9016
> >>>
> >>>
> >>>  Original Message
> 
> >>> Subject: Re: [AFMUG] B11
> >>>
> From: "Jaime Fink" 
> >>> Date: Tue, August 2, 2016 4:49 pm
> >>> To: "SmarterBroadband" 
> >>> "af@afmug.com" 
> >>> 
> --
> >>>
> >>> > The PHY is 1733 Mbps aggregate, so depending on the chosen window
> >>> sizes, the top real world TCP speeds we’ve seen are between 1200-1300
> Mbps
> >>> aggregate (75/25 or 50/50 mode), or 600-650 Mbps symmetric (50/50).
> >>> >
> >>> > For 11 GHz this assumes high/low pairing of 80 MHz and both
> >>> polarizations coordinated.
> >>> >
> >>> > Jaime Fink • Mimosa • CPO & Co-Founder
> >>> >
> >>> >
> >>> > On August 2, 2016 at 1:34:31 PM, SmarterBroadband (
> >>> li...@smarterbroadband.com) wrote:
> >>> > Can the B11 do 1Gbps Symmetrical?
> >>> > If not what is best Symmetrical?
> >>> > Thanks
> >>> >
> >>> >
> >>> >
> >>

Re: [AFMUG] B11

[John Blake]

All in for a Huawei RTN-380 including brackets, antennas, software,
licensing, cables, accessories, etc. comes out between about $22k-30k per
link depending on which options you pick.

Also, if 4Gbps isn't good enough, there is the RTN-380H that does 10G
wireless.  The specs on this are insane.

   - 10Gbps throughput
   - 2000 Mhz channel spacing, 128QAM modulation
   - SFP+ interfaces


On Wed, Aug 10, 2016 at 11:00 AM, Joe Novak  wrote:

> I'm talking 706 FT, give or take 6 inches
>
> On Wed, Aug 10, 2016 at 11:52 AM, Eric Kuhnke 
> wrote:
>
>> Depends on your distance, if you're going 1 km vs. 2.5 to 4 km
>>
>> It is definitely a VERY narrow beam width. Particularly with 60cm
>> antennas. I wouldn't do it on anything that sways.
>>
>>
>>
>> On Wed, Aug 10, 2016 at 6:07 AM, Joe Novak  wrote:
>>
>>> How stable of a structure do you need for 70/80Ghz? Self standing rohn
>>> 45 @ 55 ft too much? It's bracketed at 25ft.
>>>
>>> On Wed, Aug 10, 2016 at 7:02 AM, Kurt Fankhauser <
>>> lists.wavel...@gmail.com> wrote:
>>>
>>>> What price range is a Huawei link in?
>>>>
>>>> On Wed, Aug 10, 2016 at 2:47 AM, Eric Kuhnke 
>>>> wrote:
>>>>
>>>>> You cannot directly compare 11 to 80 GHz. Totally different market. I
>>>>> can't do beyond 2.5 km at real five to six nines uptime with 80 GHz (even
>>>>> with +18 Tx power radios).
>>>>>
>>>>> I can do 60 km with 11 if the link will tolerate some ACM.
>>>>>
>>>>> On Aug 9, 2016 9:31 PM, "John Blake"  wrote:
>>>>>
>>>>>> It's obviously pricier than the B11, but if you want true symmetrical
>>>>>> and a ton of throughput, you could look at the Huawei RTN-380 radios.
>>>>>> These will do 4Gbps (2Gbps symmetrical) at full licensing, or there are 
>>>>>> 1,
>>>>>> 2, 3Gbps licensing options.  They use 71-76 GHz and 81-86 GHz and so are
>>>>>> super easy and cheap to get (lightly) licensed. These are very popular 
>>>>>> with
>>>>>> carriers outside the US, but not so much in the US because of all the
>>>>>> Huawei/Ciscolobby disputes, but that has largely been resolved and they
>>>>>> have been getting traction here.  Let me know if you want more info, we 
>>>>>> are
>>>>>> Huawei VAR.
>>>>>>
>>>>>> John
>>>>>>
>>>>>>
>>>>>> On Wed, Aug 3, 2016 at 9:54 PM,  wrote:
>>>>>>
>>>>>>> Gino,
>>>>>>>
>>>>>>> That seems to be the only configuration that meets the requirement.
>>>>>>> The way I see it, in a traditional FDD system you would license an XPIC
>>>>>>> pair of frequencies, say 11075 H/V at point A and 11565 H/V at point B.
>>>>>>> Assuming an 820 running 256QAM you might get 500mbps per polarity per
>>>>>>> direction, for a two-way aggregate of 2gpbs.  You would have licensed
>>>>>>> 160MHz at each end of the system, for a system efficiency of 
>>>>>>> 2gbps/320MHz =
>>>>>>> 6 b/s/hz.
>>>>>>> If you operate a B11 on the same pair of frequencies, you have to
>>>>>>> use their FD mode which a typical PCN claims will deliver 736mbps.  It
>>>>>>> would seem that this can be viewed as a two-way aggregate rate, because 
>>>>>>> the
>>>>>>> radios still take turns transmitting as in a true TDMA system.  So, the
>>>>>>> system efficiency here is 736mbps/320MHz = 2.3 b/s/hz -- below the
>>>>>>> requirement of 3 b/s/hz.
>>>>>>>
>>>>>>> If you switch the B11 to the normal TDMA mode you will need to
>>>>>>> transmit on the same frequency from each end of the link.  So, in 
>>>>>>> addition
>>>>>>> to licensing 11075 H and V transmitting from point A, you also need to
>>>>>>> license 11075 H and V transmitting from point B, which adds another 
>>>>>>> 160MHz
>>>>>>> at each end. The PCNs show this configuration giving 1.47gbps (again
>>>>>>> assumed to be an aggregate figure due to the TDMA mode).  This is an
>>>>>>> efficiency of 1.47gbps/640MHz = 2.3 b/

Re: [AFMUG] B11

[John Blake]

But that B in Shannon's Theorem keeps getting big.  That 10Gbps radio uses
2Ghz channels.  That results in a lot of C

On Thu, Aug 11, 2016 at 9:02 AM, Ken Hohhof  wrote:

> There are bumper stickers “Obey Gravity ... It’s the Law”.
> Maybe you need to sell Shannon’s Law bumper stickers.
> Ooops, it’s Shannon’s THEOREM.
> So you can be a Shannon denier.
>
>
> *From:* Chuck McCown 
> *Sent:* Thursday, August 11, 2016 9:36 AM
> *To:* af@afmug.com
> *Subject:* Re: [AFMUG] B11
>
> When they figure a way around this let me know:
>
> C=B Log2 ((1+S/N)
>
> (Shannon/Hartley)
>
> *From:* Kurt Fankhauser 
> *Sent:* Wednesday, August 10, 2016 8:24 PM
> *To:* af@afmug.com
> *Subject:* Re: [AFMUG] B11
>
> wow 10Gbps over wireless? Give it a couple years and there will be an
> 80ghz Airfiber doing these speeds I can't wait.
>
> On Wed, Aug 10, 2016 at 3:10 PM, John Blake  wrote:
>
>> All in for a Huawei RTN-380 including brackets, antennas, software,
>> licensing, cables, accessories, etc. comes out between about $22k-30k per
>> link depending on which options you pick.
>>
>> Also, if 4Gbps isn't good enough, there is the RTN-380H that does 10G
>> wireless.  The specs on this are insane.
>>
>>- 10Gbps throughput
>>- 2000 Mhz channel spacing, 128QAM modulation
>>- SFP+ interfaces
>>
>>
>> On Wed, Aug 10, 2016 at 11:00 AM, Joe Novak  wrote:
>>
>>> I'm talking 706 FT, give or take 6 inches
>>>
>>> On Wed, Aug 10, 2016 at 11:52 AM, Eric Kuhnke 
>>> wrote:
>>>
>>>> Depends on your distance, if you're going 1 km vs. 2.5 to 4 km
>>>>
>>>> It is definitely a VERY narrow beam width. Particularly with 60cm
>>>> antennas. I wouldn't do it on anything that sways.
>>>>
>>>>
>>>>
>>>> On Wed, Aug 10, 2016 at 6:07 AM, Joe Novak  wrote:
>>>>
>>>>> How stable of a structure do you need for 70/80Ghz? Self standing rohn
>>>>> 45 @ 55 ft too much? It's bracketed at 25ft.
>>>>>
>>>>> On Wed, Aug 10, 2016 at 7:02 AM, Kurt Fankhauser <
>>>>> lists.wavel...@gmail.com> wrote:
>>>>>
>>>>>> What price range is a Huawei link in?
>>>>>>
>>>>>> On Wed, Aug 10, 2016 at 2:47 AM, Eric Kuhnke 
>>>>>> wrote:
>>>>>>
>>>>>>> You cannot directly compare 11 to 80 GHz. Totally different market.
>>>>>>> I can't do beyond 2.5 km at real five to six nines uptime with 80 GHz 
>>>>>>> (even
>>>>>>> with +18 Tx power radios).
>>>>>>>
>>>>>>> I can do 60 km with 11 if the link will tolerate some ACM.
>>>>>>>
>>>>>>> On Aug 9, 2016 9:31 PM, "John Blake"  wrote:
>>>>>>>
>>>>>>>> It's obviously pricier than the B11, but if you want true
>>>>>>>> symmetrical and a ton of throughput, you could look at the Huawei 
>>>>>>>> RTN-380
>>>>>>>> radios.  These will do 4Gbps (2Gbps symmetrical) at full licensing, or
>>>>>>>> there are 1, 2, 3Gbps licensing options.  They use 71-76 GHz and 81-86 
>>>>>>>> GHz
>>>>>>>> and so are super easy and cheap to get (lightly) licensed. These are 
>>>>>>>> very
>>>>>>>> popular with carriers outside the US, but not so much in the US 
>>>>>>>> because of
>>>>>>>> all the Huawei/Ciscolobby disputes, but that has largely been resolved 
>>>>>>>> and
>>>>>>>> they have been getting traction here.  Let me know if you want more 
>>>>>>>> info,
>>>>>>>> we are Huawei VAR.
>>>>>>>>
>>>>>>>> John
>>>>>>>>
>>>>>>>>
>>>>>>>> On Wed, Aug 3, 2016 at 9:54 PM,  wrote:
>>>>>>>>
>>>>>>>>> Gino,
>>>>>>>>>
>>>>>>>>> That seems to be the only configuration that meets the
>>>>>>>>> requirement.  The way I see it, in a traditional FDD system you would
>>>>>>>>> license an XPIC pair of frequencies, say 11075 H/V at point A and 
>>>>>>>>> 11565 H/V
>&

Re: [AFMUG] 90 degree LC fiber connector

[John Blake]

We (N1 Networks) just got a shipment of G.657b3 bend insensitive cables.
You can bend and smash the heck out of these and they will only give you
negligible loss.  Technically they are rated to 5mm bend radius, but they
actually do much better than that.  As long as you have room that the
connector seating into the transceiver or bulkhead connector isn't
compromised, these will work.  I'll send you a price list via direct email
as we don't have them on our website yet.

John

On Mon, Jun 29, 2015 at 12:34 PM, Chuck McCown  wrote:

>   I would think that bending losses would be greater than some would find
> acceptable.
>
>  *From:* mailto:p...@believewireless.net 
> *Sent:* Monday, June 29, 2015 12:23 PM
> *To:* af@afmug.com
> *Subject:* [AFMUG] 90 degree LC fiber connector
>
>  Does anyone know of a nearly perfect 90 degree LC fiber connector? I've
> seen the ones that bend over an inch or so but I'm looking for something
> that is nearly a perfect 90 degree angle do to the confined space of the
> cabinet. Anyone know of such a thing?
>

Re: [AFMUG] Another neighborhood on fiber, sexy sexy!

[John Blake]

If you need 48 ports per 1RU, then I would recommend you consider the
Huawei S5700-52X-LI-48CS-AC.  They do need moderate cooling, but have
excellent cost-per-port and all the features you would want for Ethernet
based FTTH (downstream port isolation, per port rate-limiting, SNMP, sticky
MAC, stacking, tons of other features). We have one in our lab we have been
evaluating for another FTTH project and they are awesome.  Much better
feature set than Planet.  They use the dual-port CSFP transceivers to get
48 ports into that space.  The switches are about $2k each depending on
quantity.  We are a Huawei VAR so message me if you want more info.

John Blake
N1 Networks
jbl...@n1networks.com

On Fri, Aug 7, 2015 at 4:56 PM, Lewis Bergman 
wrote:

> I am shocked there is no writing on the sheet describing connections ;).
> It does underscore how much more gear switching takes over PON.
>
> On Fri, Aug 7, 2015, 4:35 PM Sterling Jacobson 
> wrote:
>
>> This time I used a super dense 4U LC fiber panel that can connect up to
>> 576.
>>
>> I've got about 370 of it loaded with fiber, about a third of that will be
>> connected here.
>>
>> It requires the thinner mil cable which is the only PITA about this setup
>> really.
>>
>> It's using 24 port 1U fiber switches, but I'm still looking for a good
>> Planet rep to get the 48 port 1U density.
>> I could upgrade to those and fill out the entire 12U switch space to
>> match the 576 panel capacity.
>> Not going to have 100 percent take rate I'm sure, but it's nice to know I
>> can get the density in one cabinet.
>>
>

Re: [AFMUG] Another neighborhood on fiber, sexy sexy!

[John Blake]

If you take the entire switch and divide the costs just into the number of
customer facing ports, the cost comes out to $41.66 per port.  For a switch
platform that includes all licensing costs, stacking, 10G uplink ports, and
Enterprise feature set, I have found that to be a very good price.  I would
love to know about alternatives anyone has found that competes with that.
Also, that is a single switch quantity price.  If you order in some
quantity, then that number will go down.

John

On Fri, Aug 7, 2015 at 5:55 PM, Gino Villarini  wrote:

> dude, how much is your per port cost?
>
> On Fri, Aug 7, 2015 at 7:37 PM, John Blake  wrote:
>
>> If you need 48 ports per 1RU, then I would recommend you consider the
>> Huawei S5700-52X-LI-48CS-AC.  They do need moderate cooling, but have
>> excellent cost-per-port and all the features you would want for Ethernet
>> based FTTH (downstream port isolation, per port rate-limiting, SNMP, sticky
>> MAC, stacking, tons of other features). We have one in our lab we have been
>> evaluating for another FTTH project and they are awesome.  Much better
>> feature set than Planet.  They use the dual-port CSFP transceivers to get
>> 48 ports into that space.  The switches are about $2k each depending on
>> quantity.  We are a Huawei VAR so message me if you want more info.
>>
>> John Blake
>> N1 Networks
>> jbl...@n1networks.com
>>
>> On Fri, Aug 7, 2015 at 4:56 PM, Lewis Bergman 
>> wrote:
>>
>>> I am shocked there is no writing on the sheet describing connections ;).
>>> It does underscore how much more gear switching takes over PON.
>>>
>>> On Fri, Aug 7, 2015, 4:35 PM Sterling Jacobson 
>>> wrote:
>>>
>>>> This time I used a super dense 4U LC fiber panel that can connect up to
>>>> 576.
>>>>
>>>> I've got about 370 of it loaded with fiber, about a third of that will
>>>> be connected here.
>>>>
>>>> It requires the thinner mil cable which is the only PITA about this
>>>> setup really.
>>>>
>>>> It's using 24 port 1U fiber switches, but I'm still looking for a good
>>>> Planet rep to get the 48 port 1U density.
>>>> I could upgrade to those and fill out the entire 12U switch space to
>>>> match the 576 panel capacity.
>>>> Not going to have 100 percent take rate I'm sure, but it's nice to know
>>>> I can get the density in one cabinet.
>>>>
>>>
>>
>