What type of traffic will your backbone be carrying?

Do you have a map of the valley?

Vermin,

The traffic will be mostly residential internet services--you know, web surfing, email, downloading movies, etc. We hope to have at least a couple of commercial users who will need to have higher data rates, and if we need more than E10 to service them, they can at least pay for the increase in cost at our POP.

Our hope is to be able to expand into VOIP as our user base grows. Our valley does not have cell phone coverage.

We expect to have to control the bandwidth that an individual user takes up. After all, a teenager on the internet can consume enormous bandwidth by just playing a few internet games.

Yes, we have maps of the valley. We are expecting to analyze the terrain with "Radio Mobile" software (freeware) that helps with siting repeaters to maximize coverage. Our 3-site coverage, using units high on the mountainsides over the valley, was reviewed with Radio Mobile and found to be a useful layout. However, theory does not always match reality--in our case, problems with winter access, remote power, etc.

Jon.

If you're talking 10 MB Ethernet shared (rather than switched) that might be a problem. Certainly, your equipment must support traffic policing, as well as traffic shaping if you plan to offer special service to some.

I don't see much problem with latency being produced by just the number of repeaters - this being based on a fairly small number of users. Switching circuits are pretty damn fast these days.

The only real bottleneck I can see, is as I stated - if this is 10 MB shared.

I had assumed that the connection to the POP was 10MB/s, but the wireless relay was something different. If the relay is Ethernet based, that would restrict the number of users per hop to the free address space - 30 or less if it is also acting as a repeater chain. Another issue is that, although the original Ethernet had protocol layers that would handle different grades of service, this is seldom implemented within the wireless Ethernet networks themselves. In addition, many wireless Ethernets are intended to handle very limited delays. I fear you need to go into the detail of the specific equipment.

The results will be very dependent on the details and set-up of your repeaters. Store-and-forward improves reliability but introduces latency - but this will probably not be significant compared with the general background if you can control the package size. Similarly, you should be able to utilise all the available bandwidth at the head of the valley if you size your repeaters correctly, and the bandwidth available to an individual user can depend mainly on how many are using the system simultaneously and on your sharing algorithm (i.e. position in the valley need not be too important if the system can be set up right).

I am not saying that this is the best solution. 15Km can be covered with cat 6 cable, with repeaters. It provides reasonable high speed and possibilities for extension. 1000ft sells for $123 so you need about $6000 for cable. Equipment plus labor cost extra, but from 1000 homes at $39.99 a month you can get $479,880 in one year. I will take the job!

I'll be happy as a clam at high tide if we can fully utilize the E10 POP feed. If we have enough customers to begin to bog down the E10, then thats a good thing...it means we have some cash flow!!

My question was coming from the angle that we have naysayers amongst us who claim that anything more than 3 hops will bring us down to dialup speed. 4 hops is blasphemy and beyond that I will cause the universe to stop turning.

It seems to me that the issue is more likely physical radio link setup and beyond that, software setup and tuning. Do a bad job with these and your speed will be down to a crawl. Do a good job and we will have a 30+ MB link with maybe 50 mS latency.

By the way, 1000 customers would be wonderful. But we only have about 250 people (#households=less) in the valley and expect about 70 customers if we do a good job and attract virtually everyone who might be interested in highspeed internet. I have thought about saddling up my D8 Cat and plowing a fiber link down the edge of the highway through the valley--but the paperwork to do this is enormous. We'll get there in time, as "fiber to the home" is probably where we will need to be in a few years.

Thanks,

Jon.

Number of hops should not in itself be a problem. And your local latency should indeed be relatively small provided no part of the network is overloaded. It's really more a case of the limitations (protocols and buffering) of the specific equipments you are using. I haven't used any low-end equipment, so I can't advise on detail, but MikroTik claim to provide free-to-use network analysis software for their own equipments (the Dude). I think that would be at least worth a try.

I too would like to know more about your wireless relays. Are we just talking WIFI routers or something more esoteric - ham radio kind of stuff?

Ooops...you got me with that question. For the MikroTik products, which are fairly highly thought of--cheap, reliable, and fast (and made in Latvia!) I think they are a takeoff of WIFI stuff but moving beyond that because I believe they do run proprietary protocols in links that display quite high speeds. www.mikrotik.com

The seem to specialize in some sort of "routerboard" that has PCI connectors designed for radio cards.

Jon.