From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wi0-x234.google.com (mail-wi0-x234.google.com [IPv6:2a00:1450:400c:c05::234]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority G2" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id 71CB121F2ED for ; Tue, 30 Sep 2014 15:49:33 -0700 (PDT) Received: by mail-wi0-f180.google.com with SMTP id em10so40283wid.13 for ; Tue, 30 Sep 2014 15:49:32 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; bh=BL8CHROEhqGkMums4hc/M3PS9OqRKV+SmJh63CC6G78=; b=TZPRhbIW5ypbTZxh4kvGFUmAAecc5+7EulqiEQ0F+fIEGQwwFv7t7Ois9EPOXBVVED az37sd3OkW6jIlCPYX+FyhKipK68kuGVgNONHMm4CIpvbK1ej6YMjQTSgKGkuPQlwRxP n6UNFoiW84ElSem7OdYBHapNmXohqG9437CKbW0cgSemKsLJplu3CLgwfpkPOnFUsQuM phPyeE2vJhkXr6bbiKIUKEjZLzfgJpVelwKnI7/nKqsppSNtx1QBaVBkkf++MOio6Ctz lVQlVjkXfjPfkzsKHwhWnq0I7CTNruFzJvyBymR9aszc9CUQdOq/96BUyFZ7Oyhjs4Zx HuKQ== MIME-Version: 1.0 X-Received: by 10.195.11.200 with SMTP id ek8mr58575715wjd.85.1412117371732; Tue, 30 Sep 2014 15:49:31 -0700 (PDT) Received: by 10.194.172.74 with HTTP; Tue, 30 Sep 2014 15:49:31 -0700 (PDT) In-Reply-To: <20140925145342.GV15839@angus.ind.WPI.EDU> References: <20140925145342.GV15839@angus.ind.WPI.EDU> Date: Tue, 30 Sep 2014 15:49:31 -0700 Message-ID: From: Dave Taht To: Dave Taht , "cerowrt-devel@lists.bufferbloat.net" , Charlie Perkins Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Cerowrt-devel] Fiber to the yurt approved X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 30 Sep 2014 22:50:02 -0000 On Thu, Sep 25, 2014 at 7:53 AM, Chuck Anderson wrote: > On Thu, Sep 25, 2014 at 07:06:30AM -0700, Dave Taht wrote: >> searching alibaba for new gear is truly an eye-opening experience... >> >> examples: >> >> http://www.alibaba.com/product-detail/Atheros-based-router-with-sfp-port= _1979360314.html >> >> http://www.alibaba.com/product-detail/Factory-OEM-Ralink-3052-300M-OpenW= rt_1923358888.html >> >> I am under the impression however, that SFP+ is needed for gig >> uplinks. Not that 100Mbit >> is bad... > > No. SFP was originally 1 gigabit (1000 mbit) over fiber only. Now it > can be 100/1000 megabit dual-mode copper or fiber or 10/100/1000 > megabit tri-mode copper or fiber. SFP+ is 10 gigabit. There are > other speeds as well, but those are used for things like Fibre > Channel, SONET, ATM, etc. rather than Ethernet. > >> On Wed, Sep 24, 2014 at 11:12 PM, Joel Wir=C4=81mu Pauling wrote: >> > Put as many pairs as you can fit into the conduit to leave quite a lot= of >> > slack (2-5metres) >> >> The specific question was what form of conduit (earth burial) would be s= ane? >> This would be a ground burial thing where I would be comfortable with >> ground buriable ethernet, no conduit (it's a forest), but fiber looks mo= re >> fragile. >> >> > We bury our splitters with ofdm break outs in waterproof boxes every 5= 00m - >> > 2km or so for the GPON roll out and blow the Fibre to the premise from= the >> > split. >> >> Residences and yurts are spaced about every 3 meters in distinct subsect= ions. >> >> > Burying splinter boxes prevents vandalism/flooding issues and >> > accidents. >> >> This is a campground in california. Amusingly, a few hours after we star= ted >> eagerly discussing digging up the ground, it started pouring rain. I'm g= oing >> to put "discuss fiber deployment" into my bag of drought-ending tricks..= . >> >> >Con's you have to dig it up every time you want to connect >> > another pair into the ofdm. Fibre blower kit is expensive. >> >> I saw that a fiber blower box was 6k. There seem to be daily rentals >> available... There are also tools less expensive than a blower like these: http://www.amazon.com/gp/product/B00K6AEBN6?pldnSite=3D1 Too tedious and unreliable? Lord knows it's hard to crimp ground burial cat6 gel.... takes me 6 tries every time, and if fiber is harder, oh, my.... >> >> > Choose your connectors on the ofdm carefully. LC style connectors are = the >> > norm on SFP(+) optics and isn't angled. This is my personal favorite b= ut >> > some dislike it. Is rarely the norm for In premise kit especially for = pon. >> > >> > Angled and unangled SC connectors are norm for PON and CPE Home kit. A= ngled >> > is better for loss and used primarily on the splitter OFDM but easy to= munge >> > if you connect angled to unangled. > > Angled (APC - Angled Polish Connector) usually has a green colored > connector, while non-angled (PC or UPC - Ultra Polished Connector) > usually has a blue connector for singlemode. You can't mate two > unlike kinds (well you can try, but the loss will be high). > > Angled improves the Return Loss figure (larger negative dB value, > usually -65 dB vs. -45 dB on regular UPC) due to it preventing > reflections at the connector mating point from entering back into the > trasmit fiber causing attenuation. Instead it causes the reflections > to divert to the side. IME, it is used for PON and HFC networks, or > any network that sends a RF modulated signal down the fiber. For > straight data-only deployments, usually un-angled is used. Definitely > all SFP or SFP+ modules you find will require the regular PC or UPC > kind, not angled. > >> > You might even consider just not using a splinter or OFDM patch at all= and >> > just having slack and unterminated fibers. Cheap Chinese Fujikura equi= valent >> > spilcers can be had for around 3000$ now. And splicing is always bette= r than >> > patching IME. >> > >> > Have fun! >> >> You just hit me with more condensed jargon than I've had to deal with in= many >> a month, but I think I grokked most of it. It has been really >> interesting to absorb >> an entirely new technology with your help, that of the list, and google. >> >> First up is just to trial something between two desparately needed point= s... > > Outside Plant (OSP) cable is designed for outdoor use (but not > necessarily direct-burial, may still need conduit). There are > regulations on how far into a building you can bring OSP cable (50 ft > or 50 meters?) due to fire codes. There is also Indoor/Outdoor cable > to get around that limitation. > > Given you are using regular routers or switches, you don't want PON > (star arrangement of fiber that passively splits to multiple > destinations in the downstream direction, and uses timeslots for the > upstream direction, requires special OLT and ONC gear at each end and > isn't really regular Ethernet). So you will generally need two > strands for each point-to-point link (although there are special > bi-directional SFPs you can get that will run TX and RX over a single > strand using WDM, not recommended due to higher cost). > > Do you have a central aggregation point planned? If so, I would > home-run all the strands from each peripheral building to that point > (or rather, splice separate cables together where necessary so that > logically each building has a home run of at LEAST 2 strands, although > I would plan on at least 6 or 12 for > expansion/redundancy/spares/experimentation). The labor is the > expensive part, so pulling as many strands as possible in one go is > the best way. You don't have to or splice all of them or terminate > all of them onto LC/UPC connectors yet. > > I can ask what our contractors recommend as far as type of conduit. --=20 Dave T=C3=A4ht https://www.bufferbloat.net/projects/make-wifi-fast