Planning a new exchange / La planification pour un nouveau commutateur

As I mentioned in my earlier posts, one of the key items in our village may be the design of the new telecommunications exchange they are building next year. I wrote to some Dutch friends (fiber experts) asking for their advice on a number of key questions.

Here is my attempt to paraphrase their answers:

1. Question:
   Would it be better to build a separate exchange for fiber next to the copper exchange or to put the ODF in the same physical building?

   Answer:
   It is extremely important that the incumbent operator can't "hold anyone hostage" if fiber is going to terminate in the same building as copper. There are a number of points to consider in such a case:

   • There needs to be a room which is separate from the main DSL equipment and that is accessible from the outside for the ODF and fiber interconnection. This room needs to be large enough to accommodate operator's active equipment if necessary.

   • The municipality needs to have the key and control access rights to this room.

   • There needs to be a conduit from the room to the perimeter of the land owned by the incumbent and this needs to be large enough to accommodate all the fiber coming in (see size questions below).

   • There needs to be a separate power supply from France Telecom - essentially a second meter for the municipality to the room.

   • The room needs to be large enough to house a UPS and air conditioning if necessary.

   • There needs to be a hole in the wall between the fiber room and the ISP's equipment which is large enough to pass 3000+ fibers (1500 households x 2 + some spare)

2. Question:
   How much space in the building/room is necessary to house an ODF?

   Answer:
   One ODF will support 800 fibers and it measures roughly 1 meter x 50 cm, including space for patch cords. You need to be able to access the ODF from the front and the back.

   One 19 inch rack will support the electronics for roughly 400 users, dissipating 3-5 watts per user. The rack needs to be accessible from the front side.

3. Question:
   In planning for conduit access, how many fibers should we plan on for each home?

   Answer:
   There is no need for more than 2 fibers per residence. The trend is actually moving to one fiber per residence with three colors per fiber (data upload, data download and television). Putting in two fibers at the same time is a good way to hedge your bets though.

4. Question:
   How much space will we need inside existing conduit to push out point-to-point fiber from the exchange point to the village?

   Answer:
   One of the projects in the Netherlands uses cables with 96 fibers inside. They can bunch 6 of these cables (6 x 96 = 576 fibers) into one 50 mm HDPE duct. It would take 6 of these 50 mm ducts to provide 2 point-to-point fiber connections to each of the 1500 households in the village, with a little to spare.

   Using these numbers, the space we'd need for the conduit leaving the exchange (before disaggregation) would be 100 mm x 300 mm.

---

As always, I'm grateful to my Dutch friends for their advice.

Two steps behind - Deux pas derrière

Talk about missed opportunities. This is the residence just next to ours which dug up their sidewalks this week - without putting in any new fiber or conduit. It is really too bad we are not better coordinated in our village because this is one of the next two residences I wanted to target for expansion of our own network.

They do have existing conduit which we could probably use but this would have been an opportune time to put in a telecommunication chamber (hand hold) as well.

I'm going to try and schedule a meeting with the mayor soon so we can get an idea when these projects are going on.

One step behind - Un pas derrière

I walked past the spot of my preferred aggregation point this morning on the way to work and saw workers getting ready to redo the surface of the parking area. It's too bad I didn't know about it because it would have been an excellent time to drop in the box.

It's not too bad though because I still don't know who owns that particular parking spot I wanted to use and whether they would have let me put the box there. Also, I started wondering how I would be able to tap into France Telecom's box which was in the middle of the public street. It may be better for me to simply move one "hop" back onto onto the communal property and put the box there.

Still, it highlights one of the difficulties of planning a network like this. It is difficult to coordinate work.

Perfect aggregation point? Or not. - Un point d'agrégation parfait? Ou pas.

I probably made a number of neighbors nervous yesterday walking around with my map and GPS in hand trying to locate hand holes in their property from the sidewalk. However, the exercise helped me decide on the area were I think it would be best to place the aggregation point. I found it at the bottom of the residence, right next to the small shops and across the street from the France Telecom trunk conduit.

I initially looked for a place to install an ODF because it allows faster access to the connections and seems to be easier to add connections if we expand outside our residence. I was interested in an ODF from Pacific Interconnections, LLC in the US (see photo). However, I quickly realized I'd take a big risk putting an outdoor ODF in an area of the residence which has lots of delivery trucks.

I'm probably more sensitive to trucks backing up in the residence because we came home from vacation this year and someone had backed into our car parked on the street. The paint from the vehicule that hit us was green and high up on the car - leaving the police to assume it was a rather large truck. The damage to our car was expensive to fix but nothing like trying to repair a damaged ODF. That's why I think it may be better to have a splice enclosure in a hand hole in the street.

As I mentioned in my last post, I think the best place would be next to hand hole 021 on my distance map. I took pictures of the spot this weekend (see below).

The road with the existing France Telecom boxes (shown to the left) belongs to the village, not our residence. That means it would be much more complicated to get permission to dig and install our own boxes. However, the land with the cars parked on it is part of the residence. This parking area is in fairly desperate need of a renovation and would be an excellent spot to pass off the residence connections into France Telecom's own conduit.

The nice thing about the hand hole in the two pictures is that it's one "hop" away from the main conduit which goes down 330 meters to the exchange. I would like to put our hand hole in front of the silver parked car in the picture.

The only catch is the parking spots are actually owned by individual tenants in the residence, not communally. That means I'd need to get permission of the owner to put in the box. I'm not quite sure how or if that would work.

The residence DOES own the road in front of the white garages (in the middle-left picture) if the parking spot doesn't work out. There are two FT hand holes there as well which could provide easy access. In that case I'd look at putting the hand hole next to waypoint number 020.

Finally, I inserted another waypoint out at the exchange to do the calculations for Part 2 of the project, going from the residence to the exchange (see below).

Topology and layout ideas / Idées pour la topologie

I had lunch with Vincent MONVILLE again this week and we talked about some of the fiber details of the project. I was interested in his estimate of installation costs (to see if I've set aside enough) but he explained that first I had to decide exactly how I was going to distribute and split off each connection. Only then would it be possible to get a better grasp on the prices.

He was right so I took out my trusty Garmin GPS III Plus and marked all the handholes specified in the France Telecom map. That allowed me to get a pretty good estimate of the distances from each handhole back to the main aggregation point for telephone lines in our residence.

Here are the numbers I came out with.

• Total length of all fiber pair segments to the handhole (handhole to aggregation point): 3934 meters
• Total length of all fiber pair segments to the house (generous estimate of home/apt to aggregation point): 4719 meters
• Average loop length to handhole: 79 meters
• Longest loop length to handhole: 220 meters
• Shortest loop length to handhole: 19 meters

• Best aggregation point: In the street next to waypoint #021

I figure the best way to distribute the 50 lines is using some sort of microconduit within the existing conduit in the street. That way the line would be protected at the split-off points. I'm not sure though so any suggestions would be greatly appreciated.

France Telecom: map of the conduit - France Telecom: Carte des fourreaux

I came back from the US this month and my neighbor had a big surprise waiting for me. He had called France Telecom to get more information about the conduit in the residence and to request a map.

He presented it to me at the door the day I got back. It's great. Now I can see where I need to put my own aggregation point (bottom right of the communal property) in order to get back to the exchange in the most efficient way.

There is, however, still some confusion over who actually owns the conduit. My neighbor was able to dig up some information from the original builder but it is information already had (Article 14: Passage de Canalisations). It is fairly clear that even if we don't own it we have a right to have the communal antenna (RF over fiber on the second pair) in the conduit.

He gave me the name of his contact at France Telecom and I'll follow up with him this week.

FTTH Council Europe - Notes and advice from equipment manufacturers

I'm here at the FTTH Council Europe meetings in Paris and it has been great because there are all sorts of equipment operators who can answer my questions about the rollout. I've been here with my day job but was able to stop by vendor booths over lunch.

• The opening session had the CEOs of Free, Neuf/Cegetel and Numericable in addition to the head of Orange's FTTH project. They discussed a lot of issues but at one point someone in the audience asked whether the operators were interested in connecting suburban areas with FTTH. The CEO of Neuf/Celegtel Michel PAULIN was honest and said, "We have no plans to go out there. It's not cost effective". I don't think the operators plan on coming anytime soon.
• I ran into the CEO of a conduit company (GM PLAST) whom I had met in Iceland a few years ago. We talked for a minute about the project and he mentioned that they may have some conduit I could buy cheaply that they couldn't sell otherwise because it has the wrong printing on it, etc. That's good news considering I don't need much. In speaking with all the conduit providers, microconduit seems to run between EUR 2-5 per meter - depending on a range of factors. That helps give me an idea of costs.
• I also spoke with Corning Cable Systems and they were very helpful about suggesting the most cost-effective way to rollout for my case. They said that microconduit is the best way to go and that I should connect everyone at the same time. They have solutions that allow operators to quickly add new subscribers but that's not what I'd need. After talking with them I see better how I should connect the fiber within the homes as well.
• Iit seems as if France Telecom has now made an offers for access to ducts in the previous few weeks. I'll look that up and see if it could apply in our case given we have space in the ducts going through the residence.
• On the other hand, there is an issue of getting OUT of the box in the driveways and into the homes. I spoke with Radius and they said that drilling out of the box to the house wouldn't be that difficult. That's also good news.
• KPH talked about something interesting in NL about raising awareness for FTTH. They gave little little flags to people who signed up to put in their lawns. It raised awareness and allowed neighbors to ask why their neighbors hadn't signed up yet. We don't really have fronts lawns but it was an interesting idea.

It has been very interesting and helpful in my planning.

How long of a loop? / Quelle longueur pour la boucle locale?

It's clear to me that one of the most challenging elements for network operators is actually mapping out the most efficient topology in terms of capacity and cost. I thought I had decided how I would lay out the network but now I'm starting to wonder if I should readjust after certain developments here in France.

One of my big struggles now is how long to make the "local loop" element of the network. I know that "local loop" typically refers to a copper twisted pair but I'm not sure exactly what you would call its replacement in FTTH-speak.

Initially I had imagined putting a concentration point in the middle of our residence (100 meters away) where I could install active equipment. Then any operator could come and patch into us to offer services at the edge of our network. However, now I'm beginning to wonder if I should plan on taking all fibers all the way back down to the exchange 600 meters away.

If I thought that I would surely stop at just laying fiber in our residence then locating the active equipment next to us would probably be a good idea. However, the situation becomes a bit more complicated if I ever decide to spread out the fixed costs of connecting to exchange by expanding the project to other subdivisions at a later date.

I'm not sure that it makes good economic sense to place active equipment in each subdivision of a small village. It's not cost effective for the equipment and it surely introduces more potential points of failure. At the same time, it may not be cost effective to run a large amount of fiber to the edges of the village several kilometers away either.

Right now this is playing out in Paris. France Telecom's PON network requires less physical space than Iliad/Free's point-to-point network. That means that Free is likely looking for property all over Paris where it can locate active equipment to terminate its fiber closer to users.

I don't need to make all these decisions now since I can wait to see how the project progresses and if other subdivisions are interested in connecting with us. One of the important decisions for now though is how big of conduit I should put in leaving our residence out to the exchange? Do I plan for something big enough to accommodate future growth - even if it will increase my up-front fixed costs?

FTTH advice from the Netherlands / Conseil de FTTH venant des Pays Bas.

Vincent MONVILLE suggested I talk to someone who actually does FTTH rollouts to get an idea of what type of network topology I should choose. Then he'd be able to help me calculate the costs of putting in the network.

I've made several contacts in the Netherlands with operators who have/are putting in their own open-access fiber networks and I contacted them with the question about network topology over the weekend.

I explained that I was originally interested in having a simple "active star" topology with the fiber aggregation point in the middle of the residence and an individual conduit to each home going out from the center like the arms of an octopus. Vincent raised some concern that this could mean I'd need much bigger trenches right next to the aggregation point because I'd have to put a large number of conduit together.

I wrote to the FTTH experts in the Netherlands and asked for ideas, explaining my concerns and looking for their suggestions. They responded quickly and said that running individual conduit to each house from the center was probably the best way to proceed - given our particular geographic and cost factors. The maximum number of conduit I'd be putting in any trench would likely be 8 (with the exception of the two apartment buildings). There I could probably run a larger conduit which would support a thicker cable of fibers.

I'll get back to Vincent and ask for his advice from here.

French civil costs with Vincent MONVILLE / Discussion des coûts civils avec Vincent MONVILLE

I had lunch with Vincent MONVILLE, a French fiber specialist who is a friend of Benoît FELTEN. I met him to talk about general costs and civil works here in France. He's very familiar with providing point-to-point fiber connectivity and clued me in to some of the hurdles I'm going to have to go through. I took notes and even snatched the part of the paper table cloth where he'd drawn some network diagrams.

He broke the network into three geographical parts:

1. Outside the residence
   This is one of the most complex parts of the civil works to manage bureaucratically. He suggested this is best handled by the operator from whom we eventually purchase connectivity or bring in to provide their own services. In these areas we're also required to dig down the full 60 cm for the trenches.
2. Public areas of the residence (sidewalks, roads. aggregation point)
   Vincent explained that in a residence I'll probably need to put my request to use the public spaces (and dig the road) to a vote of the residents. That seems natural. I then asked him if I'd need unanimous support from the residents. This could be difficult to obtain after watching some of the debates at the previous association meeting. However, he explained that the process for getting approval is different among associations. In some you only need permission of the "syndic" or president of the association. Others may require a majority vote. Finally, there are some circumstances when we'd have to have all residents in agreement. I'll have to check more on how this works for our own residence. One positive element - we may be able to reduce the depth of the trenches to 15 cm since the residence is private property. That will reduce costs.
3. Personal property
   Obviously here we'd need permission of every owner before digging in their property. I think most will agree to this, particularly if they're not paying up front. However, there could still be cases where people may decide they don't need fiber and don't want their lawns and gardens touched. This poses a particular problem. In order to attract operators we need to have as many potential clients as possible. Skipping a house with the network (because an owner refuses) will be slightly less costly but also will decrease the incentive for operators. Vincent suggested a work-around here. We could simply put a small drop box in the public portion of the sidewalk in front of their home. As soon as the owner changed his/her mind we could pick up the conduit from there and bring it to the house.

Vincent also explained several steps before the contractors come in and start digging up the road/sidewalk in the residence:

1. Ask the mayor for a list of network owners
   We need the list of the network owners (electricity, water, sewer, gas, phone, cable) so we can ask them for maps of where their networks are in the area. We'd need to map our own conduit around the existing infrastructure. Vincent explained this was part of a 1991 law I'll have to dig up (so to speak).
2. Wait four weeks for answers
   The operators have four weeks to answer us with their maps. These will typically come as a PDF or as paper copies.
3. Present maps and plans to residents
   Once I've mapped out the proposed route for the network I need to present it to the residents. This would show them where we would like to dig in the public areas and on their property.
4. Notifying existing network owners of the dig
   The company doing the digging has to send a "DICT" to all network owners 10 business days before the work starts. This notifies them that there may be work near their own networks and I assume it allows them to come and ensure their networks remain untouched.

Vincent also gave me a bit of information about the rack I'd need to install and the elements that I'd need to put in it. I'll try to locate this in the public area of the residence which will serve as the aggregation point for the fiber connections.

• Enclosure: 2m x 60 cm x 60 cm
• Uninterrupted power supply
• A fiber entry point
• The actual switching/termination equipment
• Batteries

Vincent gave me lots of good advice and suggested I speak next with people who have been doing FTTH projects to see how they would plan out the network. Once I have that information he said he'd be able to give me a better picture of the costs involved.

Olivier JERPHAGNON - (Calient)

Yesterday I met with Olivier JERPHAGNON who works for Calient Networks. Dirk VAN DER WOUDE suggested we meet because of the connection between unbundling/regulatory policy and the optical switching equipment Calient produces. The gist of our discussion was about how different topologies and technologies affect the way alternative operators can access end-users. Luckily, I was also able to talk to him for a minute about this project as well.

Here are a some of the key/interesting items he brought up:

Overall

1. One of the competitive operators here in France is planning to have 60 POPs for all of Paris at 10,000 subscribers each. That certainly helps in terms of scale economies.
2. Access networks are similar to airports and harbors. If cities can promote the development of airports and harbors then they should be able to put in networks as well.
3. The more "pure" the model for structural separation, the more tax money will be needed to fund the rollout of the network.
4. VDSL and PON FTTH networks typically have a few hundred subscribers in a street cabinet compared to our 50.
5. Calient's equipment is useful for creating dynamic PON and end-run networks from a central point. Their equipment is usually located in a central office (exchange) but they are working on a smaller version for street-size cabinet use.

St Nom Fiber

1. Our project is small enough that it may be difficult to attract operators to come and provide services. He suggested that operators may only consider if they can be the sole provider.
2. He emphasized that the key to attracting an operator will be having everything in place for them to connect in.
3. One area where we may have an advantage is connecting the last 10 meters. The last few meters to reach a house and install the connection are among the most difficult and expensive for operators. We will be much more attractive to operators if, with the support of the residents, we install the network ourselves (and this last 10 meters) ourselves.
4. He urged me to approach as many operators as possible about coming in and connecting into the network.

Benoît FELTEN - (Fiberevolution)

Two work colleagues and I had lunch with Benoît FELTEN of Fiberevolution to talk about the project and fiber rollouts in general here in France. We spent a good deal of time discussing why many operators were planning PON-based networks versus end-run fiber.

Some key points of the discussion:

• First fiber in will likely solidify their position as a monopoly provider. This has huge implications for competition, network architecture and regulation.
• The 50 homes in our project are peanuts for operators here in France. The key will be using this as a high-profile test case for the business model.
• This little project could help apply pressure on local candidates running for office next year. Hopefully we can make open-access fiber one of the main campaign issues at least in the village.
• The power of Web/Telco 2.0. This came up in two ways. First, it's so easy to make connections with other like-minded people with the interactive Internet of today. Second, and Benoît pointed this out, we're likely not long from the day where you can offer a complete third-party "triple-play" offer using Joost, Gmail and Skype.

I've also asked Benoît for help locating people with fiber expertise in the area. This is going to be really important because as interested as I am personally, I need to find someone who really knows installations here in France.

No single "best practice" / Non seulement un "best practice"

I've had several discussions offline the past 2 weeks with experts who have offered their advice. What I'm starting to understand better is that there is no single "best practice" for putting in a FTTH network. All cases are different and lend themselves to different network topographies, equipment and business plans.

I knew when I started that different housing layouts would require different topologies. What I hadn't counted on was the sheer number of parameters that factor into the decision. It's not like you can just say that you want to run fiber to the home. Will that be PON or end-run? Will the deployment be arial or underground? Even if you narrow that one down you're faced with what type of underground deployment (direct-bury or in conduit, etc). Then you get into questions about what kind of fiber... The list of decisions goes on and on. Luckily I'm getting very good advice from people who want to see more fiber rollouts.

All this leaves me in a bit of a chicken/egg situation. I can't convince people to sign up for the network until I know how much it's going to cost and where it will run. However, I can't easily make many of these decisions regarding the network until I have a feel for their level of interest either (and if they'll let me dig in their yards). It's tough to show up at someone's door and ask them if they're interested if you can't give them details. That's probably exactly what I'm going to have to do though.

Network topology / Topologie du reseau

I'm very interested in getting an idea on how much this broadband rollout is going to cost, particularly for the different stages. However, I've realized that the first thing I need to do is to decide on the topology of the network.

Here's a view of our residence from the air. I've marked the outline of the area I want to cover in red. The yellow lines represent where I'm planning on putting down the fiber once the street is open. The yellow star in the middle is where I'll locate the concentrator (right next to the cable company's box). Finally, the Yvelines Numeriques network and our France Telecom Exchange are in green. (You can get an idea of relative sizes by comparing things to the tennis courts at the middle right edge of the picture).

Here are the current questions I'm working on:

1. Does it make sense to run a dedicated conduit to each home in the residence?
2. What about the 24 apartments (12 in each building)?
3. Would it be best to have the conduit run all the way to each home or simply go to a small box in the ground just beyond the sidewalk - which would then require further trenching the last few meters when they signed up?

I'll post what I find.

Help from the Netherlands / Aide du Pays Bas

Today I received an e-mail from Dirk VAN DER WOUDE who is the broadband program manager for the city of Amsterdam. He passed along some great information that I'm going to start going through after the meeting this evening. I always figured the Netherlands would be one of the best places to look for examples. Dirk's information will be a big help.

One of the initial things he sent caught my eye in particular. It's a company in Sweden called Zitius which acts as a go-between for property owners and ISPs. From my understanding they operate and manage a network that allows 14 service providers to sell to any of the 50,000 homes they pass from one central location.

To me, it's a very interesting way to position themselves as a company. Certainly I think it would be great to get access to multiple providers via one pipe. I haven't heard of anything similar though here in France. There could be a market opening in areas with open-access fiber backbones for such a company to appear and cater specifically to local rollouts.

The interview Dirk passed along is from Benoît FELTEN’s blog (Fiberevolution). I don't know Benoît yet but I'm going to make an effort to. :)

The meeting with the residence association is tonight. This is going to probably be a make-or-break moment for the next year - particularly in regards to the work on the roads.

Bill Woodcock from Packet Clearing House / Bill Woodcock de Packet Clearing House

I was lucky to be at a meeting last week with Bill Woodcock from Packet Clearing House. I don't know of anyone in the world who knows more about Internet traffic exchange and I spent all night at dinner asking him quesitons about the project. I'm an economist - not an engineer - so parts of the discussion went over my head. However, he had some excellent ideas on the network architecture.

His suggestion was to run a 48 strand fibre assembly from the point where we connect into the backbone network back into our own group of houses. He suggested we locate the midpoint in the residence to terminate the fiber connection. From there we could run copper (or fiber) for the last 100 meters to each home and still be capable of gigabit speeds. This would require someone to put the equipment in their basement - or at least at the side of their house.

He also was able to pull together some numbers based on what I said we wanted. Of course we don't know how much it will cost to patch into the backbone network but he figured of the top of his head that we'd be able to put together the network for under EUR 24,000 to reach the 24 homes. I'm going to start looking into the equipment I'd need on the ends of the connections to get a firm idea.

He also told me he has a presentation that Packet Clearing House gives about setting up larger-scale networks but that the same principles would apply to smaller networks. I wrote to him today to ask if he could pass it along.