Slashdot Mirror
Wiring a Neighborhood?
mklencke writes "I'm part of a project that is developing a small neighborhood of about 30 houses in the Netherlands with technology, durability and ecological features in mind. We are looking at centralizing the Internet, TV, phone and radio access. Options we have come across are a central satellite dish, a central subscriber line, etc. Preferably, fiber optics will be used. However, it is very difficult to have a good overview of possibilities, and fiber optics technology is apparently very expensive to implement. Have any Slashdot readers been engaged in a similar project? Do you have hints or resources on how to go about wiring our neighborhood?"
Motorola Canopy (wireless) can fit the bill for the Internet part. Very fast set up. High speed. Relatively cheap. Good coverage.
Don't dig. You'll probably hit a gas line anyway.
"If you want to improve, be content to be thought foolish and stupid." - Epictetus
you might want to check out the PON stuff by all optic. probably a little pricy for what you are doing, but it fits the bill.
http://www.alloptic.com/
"Where ever you go, there you are"
I work for a contruction company that did just this a few years back. We built a 700+ home retirement community in the US. We partnered with a local Cable TV/Internet provider. The cable company installed all the in-ground components (just a normal Cable TV network plant) and provided both Cable TV and Broadband Internet Access.
As far as telephone and radio - I am sure your local companies could prove helpful.
Coding my way to the next BSOD!
I work for a local telecom, and we have several kinds of last-mile infrastructure, including both Hybrid Fiber-Coax (HFC) and Fiber To The Home (FTTH). Either way, each home gets a Residential Service Unit (RSU), which mounts on the side of the house, which separates out the fiber or coax into cable TV, Ethernet, and POTS jacks. Fiber is extremely expensive to deploy, which is due largely to the fiber transceivers. As a last-mile medium, fiber is greatly overhyped. Hybrid Fiber-Coax, which is what most cable-modem-ready cable companies have in the ground, can support any conceivable service, including voice (RF modulated or VoIP), data, and cable, all for a much lower cost. This uses "fiber to the curb", then coax the rest of the way. Especially if you're trying to design a scalable prototype, consider HFC rather than FTTH.
Speaking personally and not for my employer.
Most phones can be powered by Power Over Ethernet. This is a most excelent way to avoid wall warts.
I remember reading an article about DIY DSL here on slashdot a long time ago. I did a simple google search and found an article about a neighbourhood in colorado. verizon wouldn't give them DSL, so they did it themselves.
here's an article about them
and this is the Ruby Ranch Internet Cooperative
i know there's also software that can do DSL with nothing but an old soundcard and two copper wires. i don't know where it is, or if it's still maintained though.
cheers, and good luck.
The starting point is to work out what you want to do ... it's nice to grab the "quick list" but without a clear spec you'll bury yourself in detail.
Communications Needs - Is there a lifeline service involved (ie, does the dialtone need to be six-nines even if the Internet access isn't?). What's the regulatory regime surrounding phones in the Netherlands? Do you want to centralise services (eg mail server, Web cache) or just access? What about power backup?
Physical layer distribution - what's the tradeoff between wireless, fibre and copper? Does mobile phone coverage enter the equation?
Interconnect - you'll want phone calls to reach the outside world - have you planned for this? How do you get permission to send calls to "traditional" carriers?
Moving signals around is easy. Understanding what you want to do before you spend the money is hard...
As someone who wired every major room with at least 2 cat5 ports upon remodeling, there's one thing I'd do differently: Run 2" plastic conduit so that I can run other types of signal wires/coax/fire optic in the future.
:P I'll eventually put up wireless, but that's a whole new bit of infrastructure to build, debug and secure.
As it is, some rooms have ALREADY maxed out all the ports (I've got a 16 port hub chained off of one of my ports, other areas I've got to run long patch cables to get to the other side of the room, etc.) I do have a central box, but it could have been located in a better spot (it's currently located at the demarc point, I should have set up a secondary distribution box on the second floor and ran everything from there.) I used two 24-port patch panels, which ended up being just enough for the layout that I used.
Another problem that cropped up was running wires for a sound system, intercom, internal phones, etc. I ran 3 cat-5 wires to every room (some rooms had more) using 2 of them for each port (theoretically I can use splitters to expand to 4 ports), and using one of the 8-conductor wires for phone service (theoretically 4 lines.) Even so, there's now a lot more stuff I'd like to run, but without punching through the drywall, I'm a bit constrained as to what I can do. Not putting in fiber was a big mistake
I installed a fiber optic network to connect up about 50 homes and 3 apartment buildings on the campus of a research institute some years ago--in a tropical, developing country. The distances from the main distribution frame (MDF, or central point) were such that copper wasn't an option. We used 4 core 50 micron fiber to each building and put two ports in place using 3M Volition gear (one upstairs and one downstairs). Each port was connected to a 3M volition switch, each of which had a single mode connection back to the research center several km away. Neither wireless nor DSL were options at the time in the country, nor are they still. Later wifi was and still is used domestically by many of the scientists. The phone system uses the (gigabit) single mode fiber link also (all fibers buried btw, in plastic pipes on gravel and with warning tapes).
Lessons: Volition v sensitive to dirt, dust and ants and needed a bit more maintenance than originally expected (switches in airconditioned MDF perfectly ok though). Fiber optic network cards for PCs were disappointing and caused problems with some machines. Gigabit fiber optic cards are still not readily available--though transceivers are. If I was doing it again I'd forget fiber to the desk and put in a transceiver (needed for laptops anyway) and run Cat5 to every room from the Light Interface Unit. Transceivers are costly and fairly susceptible to getting zapped by voltage anomalies (lots of lightning and power outages).
BTW each volition switch connects to a different data center, each with its own ISP and broadband link, and they're cross connected with load balancing hardware (Raritan Linkproof) to maximize uptime.
In the Netherlands (where I used to live) I think you must by now be able to get a good DSL solution (it was all ISDN when I left). Do a search on DSL and multi-occupancy and you'll find some suppliers. I would TENDER FOR A SOLUTION and include maintenance of any central equipment and get a good service level agreement. You don't want to have a de facto IT dept for 30 homes.
Search on CEDIA. You may find companies that can offer a solution or help you develop a statement of requirements. Veel geluk!
Preferably, fiber optics will be used. However, it is very difficult to have a good overview of possibilities, and fiber optics technology is apparently very expensive to implement.
i .com/fsoalliance/
You can ditch the fiber but keep the optics. Free Space Optics (FSO) has been around for a long time. Despite being somewhat obscure, it is a very mature technology with a lot of things going for it. It provides fiber level bandwidth without the cost of digging up the ground to lay down fiber. Rapid deployment and high mobility can save not only money but time as well. You didn't mention how far apart the residences are in the neighborhood, but unless you're rural and very spread out, FSO may perform satisfactorily with allowance for bad weather. Bad weather being fog and scintillation.
Fog is a problem if you're near the coast or a large body of water that can produce a lot of mist. A heavy mist can really hammer the signal by several dBs over long distances on the order of a mile/kilometer. Currently it is the largest obstacle faced by permament/semi-permanent FSO implementations. Atmospheric scintillation is the phenomenon that makes stars twinkle at night. It is caused by variations in atmospheric temperature that change the index of refraction an optical signal encounters as it zooms to its destination. This problem, however, is more or less solved by making the signal take parallel paths to the reciever.
you may be interested in the following companies among others.
tellaire
terabeam
fsona
airfiber
lightpointe
industry news and references:
http://www.freespaceoptics.org/
http://www.wca
Stay sentient. Don't drink bad milk.
Google IT: "low power" "spread spectrum" "Line of sight" microwave
Line-of-Sight (LOS) Wireless, WiFi, 802.11x,
http://www.commweb.com/showArticle.jhtml?ar
http://www.ieee80
NOTE: Avoid using omni directional antennas, consider per-session (call, email, download,
EUROPE RACES AHEAD OF USA IN TECHNOLOGY!
This is a romantic return of technology to ITS place in our future.
http://www.zeppelinfan.de/html-seiten/de
http://www.flug-revue.rotor.com/FRheft/FRH
Aerial platforms provide an ideal way of delivering broadband communications services. They can be considered as a hybrid technology, combining the best of terrestrial and satellite delivery. High altitude platforms - either solar powered airships or planes, typically flying at 17-22km. no delay, no bad weather,
I hope this helps - OldHawk777
Unaccountable leaders are masters, and unrepresented people are slaves. How do US and EU fare?
In all, Holland is a great country
No it isnt, cause Holland is not a country. Holland is a province (well 2 provinces, north and south holland), the country is the Netherlands.
Eg, Philips do _not_ have their headquarters in Holland (it's in Eindhoven, province of Limburg), the Dutch TT is _not_ in Holland (it's in Drenthe), Utrecht is _not_ in Holland, etc..
Holland => a province
the Netherlands => the country
I use Friend/Foe + mod-point modifiers as a karma/reputation system.
Rick at Cable System Services. www.cablesystemservices.com They're an excellent small to midsize company that could provide anything from a turnkey system down to any individual step along the way - walkout, strand draft, design, construction, and/or testing. They also do consulting and they're world wide. I used to work there (Eagan, MN)until I moved my family back east (NY) to my old stomping grounds - I helped work on a number of projects just like yours during my time there.
s/2" plastic conduit/EMT
Some areas have strict codes against plastic conduit and/or PVC, it'll help lessen any EMI, plus it's very difficult to bend PVC. You should also substitute one 2" conduit with 2 or 3 smaller ones so you can seperate noisey conductors (speaker wire) from EMI sensitive conductors (Cat5), and reduce the cost of tools. A hand operated 3/4" EMT bender is much cheaper than a hydraulic 2" bender.
#include "futurama_joke.h";
I've been wondering why there isn't a good "do it yourself community fiber network" howto, at least as far as I know. I've been experimenting with fiber as a hobby for the last few months just for fun. Its amazing what you can find on ebay.
A few observations:
Single mode is vastly superior to multimode in terms of both range (20-100 km vs 2km) and theoretical maximum throughput (terabits per second versus gigabits per second).
On the other hand, single mode is harder to work with, and the end point equipment is more expensive. A new 100mbps fiber-to-copper ethernet converter is about $150 for MM and about $250 for SM. Gigabit gear is more expensive, but not terribly prohibitive.
The fiber itself is cheap (one article I read indicated that the wholesale cost of SM fiber is about $15 a km). All the cladding and armor they put on it makes it expensive (A dollar or two per foot for direct-bury cable with a dozen or so fiber strands).
It takes 2 fibers to make a connection, usually. This is called duplex.
It is possible to run multiple connections on different wavelengths. This is called wavelength division multiplexing. DWDM systems sometimes have over a hundred separate channels.
I don't know much about durability.
Ethernet over fiber may be good for small networks, but it requires active electronics (and reliable power) at each junction. Depending on application, this may not be a problem.
Take a look at fiberdyne's webpage if you're curious about approximate equipment costs. They seem to sell almost everything related to fiber. Here's another page with a decent fiber tutorial.
-jim
If we actually had info on the physical topology of your community it would make things easier.
But, here's what I would suggest:
Each house should have multiple pairs of dry copper running to the SAME CO. You can probably use this as your physical medium for all 30+ homes, using VDSL or "HDSL T1/E1 Modems" (ADC Makes these).
I'd find cheap VDSL ethernet Bridge/modem (which is what they are anyways) setting up one in each residence, then you can find rack mount vdsl "concentrators" or chassis which mux all of these together and give you a few ethernet ports for uplink purposes. Either that or you can use use another vdsl modem on the telephone CO side and connect them all to a standard switch (a cheap cisco 3548-XL, or a bunch of cheap 16port switches uplinked to eachother).
tut systems makes these (which ived used in the setup i've descibed) http://www.videotele.com/index.cfm Note that there's actually a bunch of competition in this VDSL (last mile) market and prices are always fluctuating. I've found single tut vdsl modems (good for hundreds of feet, 1.5mbits over a pair of copper) go for 20 bucks a peice.
I would advise against 2.4ghz wireless as it sucks. Just trust me on this. Anyone who's recommending setting up a Metro LAN on this is talking out of their ass and doesn't realize how shitty this would be (i've seen it, CDMA collisions out the ass, 200pps limit for the whole friggin network, all of your traffic cleartext, one user with the right equipment can shut it down, lmr200 or 400 cable is expensive, 2.4 sucks thanks.)
Keep to the KISS rule, use cheap CAT5/6 or pre-existing infrastucture if at all possible.
I think I managed to get it all for about $1400. The most expensive items was the cable (about $500-600 at the time), and the tester/probe kits. I didn't want to use crappy tools, so I spent about $200 on a set of el-cheapo blinking-light cat-5 contiunity testers, and a more generic telecom tester/inductive probe (since I was running telephone also.) I also spent maybe another hundred on crimpers, punchdown tools, etc. Everything else, from the keystone jacks to faceplates were either bargain-basement Belkin or off of eBay.
So, for the cost of one computer (at the time, probably a Pentium III 6000), I put in a network that is capable of supplying every room in the house (including one bathroom, the kitchen, the laundry room, and the garage, in addition to all your standard living areas) with phone and cat-5 ethernet. The $1400 did not include my labor, nor any additional equipment (ie, hubs, switches, etc.) For the switches, I have two 10/100 units (also off of eBay). I got them cheap, and they work nicely - however, I'd get consumer ethernet equipment in the future - the stuff I have right now is noisy as hell (fans) and they eat up a lot of power.
There is a problem with having this many ports though. If you have that many computers, you're going to need a lot of UPS units, and they all draw a LOT of power...
We investigated Lightpointe's solution between 2 buildings at about 150 yards apart.
Bottom line... at an excess of $35K USD for a pair, we decided to dig and run fiber at a cost of about $12K.
Ronja. The GPL'd Free Space Optical link with the $100 price tag
I think there a couple of things that need to be addressed before one starts suggesting ideas on how to set this up.
Is this going to be 30 single family detached homes or is this going to be 30 housing units in one building? Do you need a 100% fool proof network with room to grow? Or can a virus infestation that crashes your switches and routers be tolerated? What services are you going to provide? I would asume at the minmum VoIP and Data. You may also want to provide VOD and PPV TV as well as regular cable tv services. Also is there a chance the system will need to be expanded for more units, are there chances of being able to spread those capital costs with more and more users?
This is a new install. There is no point in suggesting DSL or ISDN or wireless from the last mile. You need around 100mbs to each home to provide quality VOIP, Cable TV, and Data. You will need some way to control the amount of traffic on each service. You will also need to provide some way of bundling this information together. Installing fiber is cheaper over installing copper in a new install to begin with. The part that may be expensive is the capital infustructure. Regardless of wether you using fiber or coper this will be expensive. You will need expensive switches and some way to bundle all the services into one pipe.
You may want to do some investigation into the technology that Yahoo Japan employs for there broadband system.
In the dwellings you will want to have 2-3 cat 5 outlets. 1 for phone, 1 for TV, 1 for data. You may also want to have each dwelling have its own private internal network as well as such you may want to put 4 cat connections in each room and set up one connection to goto a switch that has its own vlan for the unit. You are going to have to proabbly include a requirement that on the sale of each dwelling the ocupants are bound to a contract of 10 years for the service. They will also have to lease most of the equipment as well over that period. So you can control and re-coop the costs. Oviously this system will want to operate as a non-profit group but be-able to invest in infustructure over the life of the system 10-50 years.
You will also need to take into account support and ongoing maintance. You should proabbly hire a company that does this and is bound to provide you with a minimum service level on maintance and support. Look at a company like this one to get you up and running and look after your network: Empower
Take it from a Canadian on Broadband.
First of all, I must congratulate you on your ambitions! It is too bad that more people are not interested in developing and living in intentional communities such as the one in which you are now involved. Having said that, the most important thing for you to do, before you even begin to consider the technology that you will use to deliver various forms of information and telecommunications is to define the requirements that your solution, whatever it is, will be judged against. This will require you to develop consensus with all residents on issues like ownership of the system, maintenance responsibilities, initial capital budget and on going operational expenses. Realize that 30 subscribers is probably too many to be served on an "ad hoc" basis (unless everyone is fairly technically literate and is willing to share "on-call" responsibility) and too few to warrant a dedicated 24x7 network operations center.
Given this situation, you may find it helpful to contract out some or all of the design, build, operate, and maintain phases of the project to an established ISP in the Netherlands(xs4all, as it is owned by KPN, may or may not be interested in such an arrangement, you may have more luck with smaller competitive providers - like Demon, or perhaps even larger players like Versatel or Wanadoo). Another approach may be to "rent a NOC." This is not as far fetched as it sounds, when I lived in the Netherlands a few years back, there were quite a few experienced network engineers who took on part-time or on-call projects for extra cash. If you go this route, it would actually be better to contract with people outside of the community, so that they could monitor the state of your network from an external perspective. You could contact a company like BaseN, for example, if you would like to try this approach.
As far as the design itself, you will probably find that the monthly operating expense will be higher than if each home individually subscribed to a complement of traditional services, at least in their most basic form. Of course, you will be able to offer additional services that are either prohibitively costly for individuals, or even unavailable to consumers. You can craft a set of applications that could be specifically targeted to the needs of your community (rsync'ed gentoo mirrors, community IM and location services, VoIP centrex services, and so on). These services will help form the basis of the requirements document mentioned above.
Next, I would work from the external connectivity to the distribution of those resources. I assume that Internet connectivity is important to you and your community. If this is the case, then the connectivity requirements are probably on the order of at least 4Mbps. If network availability is a major concern then this could be provided by 2 E1 lines, diversely routed from different providers, probably to different locations within your community. It is important to realize here that this implies that your community would become in effect an ISP. If you are really ambitious, you could pull a circuit into the Amsterdam Internet Exchange and attempt to find some providers willing to peer with you.
Now, to answer the question that you originally asked, there are a number of different technologies that you can use to distribute your network. As with everything else, there are tradeoffs involved; that is, there is no right answer (but many "wrong" answers). IMHO, the best tradeoffs would be found in xDSL technology for two reasons. First, it is a relatively proven and mature technology. Second, it works over just about any reasonable twisted pair technology. On the other hand, if you will be distributing television as well, you may find Euro-DOCSIS cable modem technology appealing. This will allow you to make use of the coax infrastructure that you will need to install anyway, but you may find that CMTS (cable modem termination systems) are initially mor
You might get away with CAT 5e for the in-house wiring, but you need to consider CAT6 or multi-mode fiber for connections between switches. I just ren into this recently:
CAT 5 - 10Mhz Ethernet
CAT 5e - 100Mhz Ethernet
CAT 6 - Gigabit Ethernet
And don't scoff at the idea of gigabit inside the dwellings, either. I saw a 5 port gigabit switch for (i think) about 89.95 (US) the other day.
If you run gigabit-capable (CAT 6 or Multi-mode Fiber) from dwellings to central switch, then the resident can put in a 10/100 switch switch with a Gigabit uplink or a regular gigabit switch. Of course, there is going to be bottle-necking at the central switch, but if you put in a few OC3's maybe no one will notice...
At any rate, this sounds like a fun project.
"The Internet is made of cats."
A copy of the electrical code can be had at most electrical supply houses, and I would reccomend a copy of Ugly's Electical References. Invaluable for solving electrical wiring, conduit bending, and conduit wire fill type problems.
"The Internet is made of cats."