Outdoor Wireless Antennas

If you are trying to network the 3 buildings together what you will need is a dlink that can act as a bridge such as the 2100ap.

Then an omni in the middle and directional on the other two would work great. Buy antennas made for outdoors. Rain is not a big hindrance. Ice is worst!

Reply to
Airhead
Loading thread data ...

Omni antennas suck for such bridges. From a rooftop, you end up picking up everything in the area as interference. Better to keep the signal beamwidth narrow. At such short ranges, you don't really need the gain, but the interference rejection is important.

The DI-624+ will NOT act as a transparent bridge. As a minimum, I suggest DWL-2100AP, WAP54G, or similar box. However, you're probably going to have a big problem. These cheapo transparent bridges will only bridge 32 MAC addresses. That's probably not enough if you're planning to band-aid three entire buildings together. You'll need something that will do thousands. I suggest you look into various Proxim, Alvarion, YDI, etc products specifically designed for bridging. (Note: You will not enjoy the prices).

What kind of bandwidth were you expecting? If it's above commodity wireless data rates, have you looked into FSO (free space optics)? At only 30 meters, you can use the cheaper LED type systems that are good to several hundred mbits/sec. If you have fog, forget it, but otherwise, it's a good interference free solution.

formatting link

Like a sponge. Water is the enemy because at 2.4GHz, water literally absorbs RF. I mean, that's why 2.4Ghz was selected for microwave ovens. Spend the money on something decent so you won't be cleaning antennas in a storm.

Again, it depends on interference. From a metropolitan rooftop, I guarantee you will have an interference problem, even in the smallest cities. Put some effort into using directional antennas. If you wanna do it right, the one in the middle should have two radios and two direction antennas. Point to multipoint bridges will work but will constipate the traffic as everything will need to be store and foreward repeated from the middle bridge point. Things work best with individual links between individual buildings which really means two radios and two antennas on each building. It all depends on how much traffic you expect. If it's low, then point to multipoint will work. If it's high, individual links may be required.

If the buildings are truely inline, then you may have a clearance issue. If the line of sight between the end buildings is blocked by the middle building, then you may end up using the middle building as a repeater. This isn't all that bad, but needs to be properly engineered.

I suggest you find someone local that has had some experience engineering telco bypass and WLAN bridges, and ask them to help.

Reply to
Jeff Liebermann

10m should be no problem for the bridge, but what are you going to be communicating with? Will the other end be outside or in a building? An external antenna would likely give you more than sufficient signal. A cantenna or maybe a reflector may be in order.
Reply to
Rôgêr

Certainly. In fact, that's the preferred method for glueing large numbers of different customers together. I didn't want to get into implimentation, just take some pot shots at using cheapo wireless bridges.

Actually, even a bridge that can only do 32 MAC addresses will sorta work. Traffic on both sides of the bridge will rapidly fill up the MAC address to port bridging table. Most bridges are smart enough to expire old addresses. Some are even smart enough to push out old addresses when a new one appears. I have a custom packet generator program (not for distribution) that will belch large numbers of unique MAC addresses for testing bridges. The better bridges (usually managed with SNMP) can handle it with ease because they timestamp their bridging table and will expire the least recently used entry. Commodity bridge suck by comparison. A pair of DWL-900AP+ boxes, in transparent bridging mode, take about 10 seconds to lock up solid with an excessive number table entries. The first 32 pass, and then everything just sits there until the table entries expire. I vaguely recall that was about 3 minutes. Then, I can hit it with the next 32 MAC addresses. In practice, on small networks, this isn't all that bad as the most commonly used devices will eventually force the scarcely used entries out of the table.

Using routers instead of bridging has some really big advantages.

  1. It eliminates the MAC bridging table limit problem as it only uses one MAC address per radio link.
  2. Because it only uses one MAC address, there is no need for an overpriced wireless bridge pair of radios that use proprietary bridging protocols. A cheapo access point and ordinary client radios can be used.
  3. Multiple subnets and static routes can be used to keep unrelated customer seperate (without resorting to a VLAN).
  4. Bridging can be simulated using a VPN router on both ends.
  5. Bandwidth management by IP is much much much much much easier by IP than by MAC address.
  6. The customer can still get routeable IP addresses by first delivering an RFC-1918 private addresse, and then redirecting traffic from an incoming routeable IP to the private address. Most wireless ISP's charge extra for this. I can't seem to recall the exact Cisco IOS incantation, but can dig it out if necessary.
  7. The customer can still use NAT on their router to connect multiple computahs.
  8. Multiple connections to the internet, through multiple border routers is much easier, offering improved reliability.
  9. There's far more control over traffic and services at the IP level than at the MAC address level.
  10. Whatever else I forgot.

So, with all these advantages, why do people do bridging? Well, it's easier, much simpler, cheaper, and good enough for most applications.

Anyway, the real unanswered question is what duz the original poster plan to do with the wireless links? VoIP, video, internet, telco bypass, sell bandwidth, play WISP, private WLAN, build a metro LAN, fiber replacement, broadcast, etc? At the bottom end, almost anything can be made to work. When the number of client radios, customers, or traffic become large, solutions tend to be far more complex.

Reply to
Jeff Liebermann

James - to avoid the use of "D-Link crap:, or other lower cost alternatives that work reasonably well, why not just upgrade to the Apple Extreme product

formatting link
It allows an external antenna.

Reply to
Bob Alston

If there are more than 32 devices in the buildings, a switch is going to be needed for connectivity anyway. Can't that be a router, providing a subnet per building, reducing the traffic on the wireless that would be present if it were a flat bridged network?

Reply to
dold

I have been thinking of setting up 3 Dlink DI-624+ with 8db omni outdoor antennas to connect to buildings as it is a cheaper option than fibre and copper is not permitted (policy crap).

The wireless is solely for bridging the buildings which are no more than 15m between them.

  1. What are the outdoor antennas like in rain?

  1. The 3 buildings are inline. Would it make any difference to use directional antennas on the 2 outer buildings to point at the omni antenna on the middle building (If possible)?

Thanks, James.

Reply to
James Collins

Thanks, its also the same price!

Since we are in Northern Australia, it is hard enough finding ice in the freezer!

If I placed the unit in a waterproof plastic box (that is 2mm thick) on the outside wall under the roof overhang, would the antenna on the bridge be sufficient if the units where no more than 10m apart? (if I forget about the third building for now)

Reply to
James Collins

Oops. Comparing apples to ...

Aha! There is a switch already. That is part of Jeff's concern. But you say "a 3com switch". Is there one switch in each building?

I picture three buildings with a switch each, and wireless links between them. Most switches will relieve quite a bit of the traffic between buildings that is targeted for specific other systems, but theer might be a lot of general broadcasting on the network that will get repeated between the buildings. If these were routers, the wireless portion would be even less cluttered, and only pass the traffic that needs to pass.

If what you have already works, that eliminates a lot of the concerns that Jeff is trying to hide. He is overengineering so that you won't be disappointed, but you already have empirical test results from the existing setup.

Reply to
dold

Aack! Not 2100APs!

I have attempted to configure D-Link 2100APs as a point-to-point bridge, and they just don't work. There was another person on this group who had the same difficulties.

Essentially they would work ... occasionally ... and then the link would die for an unpredictable period of time. For some reason, the link needs *continuous* traffic to stay up.

I kludged around it by having a server on one side continuously ping a PC on the other side of the bridge.

The other guy kludged around it by streaming audio across the link continuously, IIRC. He eventually solved the problem by using a couple of LinkSys APs.

I know it wasn't a faulty 2100AP since I have several, and any combination exhibited the same behavior. Range wasn't an issue - the problem manifests exactly the same whether the APs are sitting 5 feet from each other, or 2000 feet apart with directional antennas pointed at each other.

Tech support had nothing to offer. 2100APs simply do not work as advertised in bridging mode.

As access points, they're OK. Somewhat flakier than the 3Com 7250s we have (but 1/3 the price, so I guess you get what you pay for).

Reply to
MonkeyOmen

I find it hard to believe that Dlink would make such crap. We currently do this with Apple Airports, however these are inside the buildings without external antennas. The move to Dlink is to increase the bandwidth fron 11 to

108MBps. There are about 30 PC's per building, all connect to a 3Com switch then the switch connects to the current Airport.

The bandwidth of 54 - 108MBps is suffecient.

Again, I find it difficult that we can do the same setup using Apple Airports with there builtin omni antennas, we are just looking to increase the 11Mbps to 108MBps by upgrading our equipment and possibly external antennas to decrease interference.

If I need to call in a telco then this will start to cost big $$$ so we would just spend the extra and use fibre. At the moment we are finding the wireless to be 1/4 less then the fibre alternative.

Reply to
James Collins

The 2100AP's have buggy hardware/software. DLINK acknowledges it but wont fix the problem. They operate with an intermittent signal strength that keeps dropping out. Pieces of crap.

Reply to
Q&A

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.