We think it makes sense for cost reduction for semi rural or suburban aerial distribution- reducing the fiber count to like. 12. Reduces costs dramatically vs say a 288 count and all the splicing. (Ribbons are better) -Ben
On Dec 11, 2018, at 10:49 AM, Owen DeLong <owen@delong.com> wrote:
On Dec 11, 2018, at 10:01 , Ben Cannon <ben@6by7.net> wrote:
Sure but I can fit quite a lot of fiber in very little space. eg an 864 is approx 1” dia.
Fan-outs can be done each floor, etc. And a single single mode strand has prodigious bandwidth available with the right optics.
Bonus: if you did this 30 years ago, you’re still good. Anything else (remember FDDI grade Multi-mode?) is not future proof IMO. Basic 9/125 Singlemode always will be.
In city wide deployments, a bit different, especially for eg residential service at economical pricing. GPON for sure has it’s place, I just don’t personally feel it’s inside a building all else being equal.
I’d actually argue that even if you’re going to do GPON on a wide distribution, the economics these days make a pretty good case for future-proofing with home-run fiber and putting your splitters and GPON gear in the same location. The link budgets turn out to be identical regardless of how far upstream the splitter is and once you dig the trench, the cost of the fiber itself is relatively cheap.
Home run means you aren’t locked into the PON topology if something better comes along in the future. It also opens up the potential to support competition (which I realize may be considered a detractor by some).
Owen
- Ben Cannon, AS15206
On Dec 11, 2018, at 9:24 AM, Jason Lixfeld <jason+nanog@lixfeld.ca> wrote:
On Dec 11, 2018, at 11:32 AM, Ben Cannon <ben@6by7.net> wrote:
Rip it out and run 9/125 SMF fiber home runs. Use BiDi SFPs to re-use your existing (likely SMF thankfully) cable plant. My opinion.
There’s only so much space in conduits, risers and ducts. At some point, scale would press this up against physical infrastructure realities depending on how far the active gear at the head end is from the subscriber.