A few GPON questions...

Ceiling tiles. There is a special ceiling tile that drops down that maintains our 24 port ONTs that do PoE+ to our Cisco WAPs. It’s most definitely isn’t what I would have chosen to do.

Hi Nick,

Yikes! That's upside down even for PON. The P is for Passive. The
whole idea is to not have powered infrastructure in inconvenient
locations because that's expensive and super hard to reliably

I think maybe you've found yourself in a "throwing good money after
bad" situation. You have the LAN equivalent of motion sensor lighting
in the restroom. Sounds great until you realize the maximum delay
setting is 15 minutes... how many dark #2s before you have to bite the
bullet and take it out?

This is going to cost you cash and productivity for as long as you
keep trying to make it work. Moreover, the combination of
inflexibility and elevated incidence of outages will consume the
occupants' productivity as well. There will never be a less expensive
time to install a classic (copper) cabling infrastructure than before
the building is occupied. This is where you dip in to the
organization's contingency funds and take the political hit.

And for the love of God, break the contract for the next building.
Even if you have to pay it out and just lose the money, save yourself
the expense of ever having to operate it. If they won't take your word
for it, get an external networking company to do an impact and cost
analysis so you can show the administration what they'll lose by
continuing this folly.

I'll bet your contractor had some questions they didn't want to
answer. If I'd sold you this bill of goods there'd be questions I
wouldn't want to answer too.

Bill Herrin

The cost analysis was already done. The cost is around 10% less. The implications are there is no redundancy, lacking capacity, costs were not factored in for BBUs on every ONT like should have been spec’d out for emergency phone lines, etc. After that was done management said to suck it up and make it work. Unfortunately we’re beyond stuck with it at this point as the second building is already under construction and the first buildings grand opening is January 1st.

The cost analysis was already done.

Costs were not factored in for BBUs on every ONT like should have been spec’d out for emergency phone lines.

These two things do not quite agree.

Update your CV - It is not your responsibility to shoulder the stress of your superiors’ bad decisions, especially if you have no room to learn from those mistakes.


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)


Fusion splicing …

I'm sure management did some kind of cost analysis, but ignored normal IT things like POE for things like phones and cameras. POE is super convenient in a LAN environment. But management doesn't know or usually care about that. I'd have asked the PON guys hey, what's your solution for POE delivery to powered devices? I bet factoring that in alone would have blown the PON cost up. I can't imagine doing anything as insane as PON in a LAN environment.

But everyone knows fiber is the best and you must eliminate all that silly copper, with helpful prodding from vendor sales. Could be sales didn't want IT involved because they knew IT will kill the deal, and some management will fall for that because these sales guys must be the experts. Problems later? The IT guy will deal with it.

I've had jobs where management refused to consult with or consider suggestions from IT. I once was part of an office move where the modular furniture vendor started asking questions about cabling was entering and port locations blah blah. They were told by management they don't need to know that and IT will just figure it out later. The vendor was like no way, they need to be involved now or we won't proceed. Then IT was brought in at the last minute, but if the furniture vendor hadn't refused to proceed the plan was literally F the IT guys and make them figure it out all the cabling over the weekend before everyone was to move in. Management like that just gets worse until you line up another job and quit.

Unfortunately management didn’t want an apple to apple comparison. They wanted a comparison to how it was spec’d out originally not how we typically deploy it now. I don’t think this is a deal breaker for my job, we have 50+ buildings left to manage, and our contractor is responsible for maintaining it for the most part for the next year at minimum. The previous management (CIO/CTO/their assistants) were pretty much fired over this fiasco. New management is much better to deal with.

Rip it out.

Where are your splitters? You can probably fix this with PoE 8-port switches and UPS-es.

Every day you wait will cost more.

You will never get 25 years out of this, I predict 6mos and then you rip it completely and put in copper/IDFs.


I remember working for this little company called EDS... Some bright spark decided that ATM to the desktop was the future (not this ethernet (or even token ring) thing) and subsequently converted several thousand head office machines to E3 or OC3 to the desktop. Hell of a thing trying to make OS2 drivers work for an OC3 card. That went very badly and the whole lot was ripped out again after a couple of years from memory.

Same thing happened in BA's shiny new office block near Heathrow back
in the 90's. ATM25 to the desktop and LANE. Total disaster.


I’ve used this guy a couple times, Use your favorite switch/POE switch and viola PON network using switches. Pretty sure it doesn’t work with Zhone… But Zhone is #88 in my list of PON vendor choices :wink:


My Adtran salesperson has told me on a few occasions that they don't recommend those things anymore and may even have discontinued them. Apparently they were never especially stable. Of course, your mileage may vary. I agree it would be an amazingly useful product if indeed it works well.

I've also been told that the EPON version they offer is fine, so who knows.

I have tested a variety of equipment as part of my FTTH enterprise. Active Ethernet is where I’m still sitting because I’m not quite happy with some of the PON hardware out there personally.

Yes active solutions provide more flexibility in one area but they are only viable in dense environments where the cost to build is already high and the fiber count is cheap comparably.

If you are within the b+ optics link budget there are options, but if you are spliced at your splitters any migration may be tricky.

I’ve been doing 2F drop to the home so I can later to technology migration as the cost variance is about 7c for 1F vs 10c/ft for 2F. It’s a bit more as those drops are the longer portion vs the backbone legs, but I can change from active to GPON or back without trouble.

It sounds like you have a typical vendor management problem where the equipment isn’t meeting your needs. Find a way to migrate to something else. Hopefully you have spare plant and budget to move to something else. If it’s homes, hopefully you can do a migration without coordination and entering the home, if it’s office campus see if you can DWDM or CWDM to get the capacity you need or there are other hardware like the UBNT OLT out there. A lot of the smaller FTTH types are also using the Huwaei hardware. Qualifying a vendor is hard and they change. I’ve spent decades working with my vendors trying to encourage them to do the right thing. The story you tell is a typical one of overworked employees without the power to fix the problems they see.

As others said I would consider a change as your business risk may be too high. This is where network operation becomes business risk mitigation.

  • Jared