Suggestions for connecting 2 buildings > 300 feet apart

10baseT is typically capable of running over 150 to 200 meters of CAT5 or better cable. It's not specified to do that but it it does. 10baseT span is limited by signal quality, not timing, and CAT5 has electrical properties enough better than those of CAT3 to You could also run 100VG, which _is_ specified to do that, if you can find the necessary hardware, and that would give you 100 Mb/sec.

However, it's a bad idea from a safety viewpoint as the two buildings very likely have different electrical grounds and so there may be a significant potential difference between them which may damage equipment and/or injure staff.

There are other alternatives. Doing it _right_ you'd run fiber between the buildings. Doing it _cheap_ if there's clear line of sight, you could use two wireless access points with high-gain directional antennas.

In article , Joseph Minckler wrote: :I had a question posed to me and I'm looking for some ideas on a :reliable/inexpensive (as possible) solution.

:We have 2 buildings about 450 feet apart separated by a field. :We need to get these buildings together on a network. I'm reading :varying opinions regarding the ability to connect at > 300 feet :assuming you run at 10 megabit with cat5 cabling.

450 feet apart suggests strongly to me that the two buildings have different electrical feeds. If so, then if you were to connect them by any kind of electricity carrying cable, you could run into substantial problems with ground differentials, and could end up frying the connecting equipment.

When you have areas that do not have a common ground and you need to put a LAN between them, you should likely be putting in fibre or wireless.

:I need a reliable link between the buildings, so if 10-baseT beyond :100 meters is not within spec and reliable, I need to go with :something else.

The traditional answer has been to install fibre in a waterproof shielding, and fibre is still a good choice if you need more than about 20 megabits per second half duplex bandwidth. But if about 4.5 megabits per second half duplex or about 22 megabits per second half duplex are acceptable, then a single 802.11b (~4.5) or

802.11g (~23) outdoor wireless bridge could be a strong contender when you have clear line of sight between the two facilities. Either 802.11b or 802.11g can -easily- go 450 feet if you add nearly any kind of external antenna on. The more directional the antenna, the stronger the signal you will get when the links are properly aligned, and thus the more link loss you would be able to withstand due to -thin- obstacles (such as the fringe of a tree's leaves) or fog or mist (rain drops are the wrong size to affect 2.4 GHz wireless but fog or mist can), or due to misalignments as the wind moves the antenna.

For more information on the possibilities with wireless, I suggest that you post to alt.internet.wireless, giving details of any potential obstructions such as trees or road traffic. 450' is not far at all for wireless, provided that you have a clear line of sight through a big enough volume of space [radio doesn't go point to point like a laser beam, so it needs a clear space around the centre of the path.]

And given the insecure nature of wireless, I strongly recommend an encrypted vpn tunnel between the wireless access points. For example a pair of Netscreen 5XPs at either end, between the access-points and the network. Plus, this would allow laptops with the proper vpn software to connect. Much better to be secure that have to deal with getting hacked by anyone driving by with a laptop.

The "insecure nature of wireless" is greatly exaggerated. Do you know of any cracks for WPA/802.11i? For that matter, do you know of any facilities using 128 bit WEP-Plus that have in the real world and not some contrived test "gotten hacked by anyone driving by with a laptop"? And that leaves aside the kind of signal that your hypothetical wardriver can get out of the side lobes of directional antennas.

Fiber and radio. Dont even think about STP/UTP

Thanks to all for your suggestions, they helped immensely!

In article , fellow wrote: :When you say don't have a common ground, you mean that the resistance of the :soil might be too "high" for them to be considered at the same potential, :right? :Even within buildings, it isn't unusual for the earth from one outlet to the :other to vary because of earth leakage and magnetic coupling, so logically, :for safety reasons, it isn't recommended to connect up different lans in :different rooms!.

Within the same building, if you are starting from the same power feed and splitting it up, then the ground potential should have only

-relatively- small variations. When you are working with different buildings, you do not immediately know if the buildings are served by the same power company, let alone the same power line, so you must prepare for the possibility of substantial ground differences.

This isn't normally a problem. The Chunnel has only 3.6V of ground differential across it. But during lightening storms, a nearby (100m) lightening hit can elevate ground potentials so that one stake is much higher (500+V) than another.

DO NOT RUN COPPER ETHERNET BETWEEN BUILDINGS.

This is called a ground-fault and can be very dangerous. The only reason for single-building grounds to differ is current leakage from hot into ground rather than neutral, plus a high-resistance ground path. Not acceptable even in the strange land of ring-mains.

-- Robert

When you say don't have a common ground, you mean that the resistance of the soil might be too "high" for them to be considered at the same potential, right? Even within buildings, it isn't unusual for the earth from one outlet to the other to vary because of earth leakage and magnetic coupling, so logically, for safety reasons, it isn't recommended to connect up different lans in different rooms!.

I presume then that you run everything encrypted on fiber. Not that that's completely secure.

It's available off the shelf from several vendors, including the Cisco consumer line. And not particularly expensive.

"Relies on" or "can use"?

Nope.

Wireless has yet to prove itself as secure, no matter how theoretically secure WPA and WPA2 are claimed to be. Do you recall how secure WEP was originally claimed to be? Know any big businesses that use it without additional protection?

Until it has a proven track record, I consider wireless to be insecure. My business network is too sensitive and valuable to risk otherwise.

Since 802.11i was only ratified less than 6-months ago, finding equipment at a reasonable price might be an issue. Properly setting it up and being sure the vendors implemented it right is another issue. You do realize it relies on RADIUS and AAA accounting which can be a bitch to setup securely. I suspect the OP might not be very IT saavy.

I also assumed the original poster was going for a cheap option and suggested vpn as a way of securing a cheap wireless implementation. Going with new WPA2 equipment might be more expensive than fiber.

-Chris

No just the wireless connections. Fiber at least needs to be touched to be sniffed and the fences and guards are good for that.

If you have fences and guards your security needs are likely a good deal higher than those of most businesses.

Perhaps. Then again, some big businesses deal with very sensitive stuff as well. Industrial espionage is at an all time high. Not so much for small mom & pop businesses, but some businesses could be devasted if sensitive stuff like business plans and designs are leaked out. Even worse, software companies need to be very sure people are getting in and making changes. Anyone want a copy of half-life2?

-Chris

In article , J. Clarke wrote: :If you have fences and guards your security needs are likely a good deal :higher than those of most businesses.

If he had said "fences and guards and guard-dogs", then I might agree with you, but:

- We have some fences to keep people from falling in the holes where we are building. We have other fences to keep people outside the intensity range of the supermagnets that has been deemed 'safe' to the unprepared public.

- Our premises (at the edge of the downtown office buildings) happen to border upon the part of our city that is commonly believed to have the greatest concentration of murders; I would evaluate the high-murder area as being at least 6 blocks away... but the area next to us

-does- appear to have the highest concentration of drug deals, so fences in the area are a common petty- theft- prevention measure>

- We have guards all night, every night -- but mostly they are there to check up on the boilers, air blowers, cryo-freezers and other similar equipment. Having a human on hand is less expensive than investing heavily into intelligent monitoring systems that can (for example) hear the "wuffing" noise that the air blowers make when the bearings start to fail, and make decisions about whether to call someone out of bed to look at them.

NICs have to be able to withstand a few hundred volts, without creating a safety hazard. If you've got that much difference between grounds in 2 nearby buildings, you've got a very severe problem, that must be corrected, before someone gets killed.

Network cards are isolated from the cable by transformers. There is no other connection between the card and cable.

The phone company runs copper cables miles long, between buildings. What makes ethernet so different?

I recently read a book called "Spy Dust", which was written by a couple of CIA agents. One operation they wrote about, was stealing an encryption machine, that was bolted down, in a locked vault, in a Soviet embassy. Even fences and guards don't do much, against a determined and capable intruder.

Incidentally, according to that book, some of the stuff the CIA uses, is right out of "Mission Impossible"!