Confusing question from me :)
I'm gonna setup a wireless connection in a few months the devices are below;
-4 x Cisco 1310 series bridge (without ant)
-4 x Cisco 21dbi Dish Antenna
-2 x Cisco 2950 series 12 port switch
2 buildings will be connected, same network. But 54mgbit half dublex is not enough for my customer.So i will use 2 cisco 1310's at each building. But one is not for redundancy it will be used at the same time. with a total of 108mgbits half dublex speed.
Cisco 1310 bridges have etherchannel feature(also link aggretiation 802.3ad supported at layer 2 i think) The same feature is also used in cisco 2950 switches, this technology is used to combineond different ports to show as a single virtual port.(also i never tried, its only theory) :)
-Is this topology ok? Can be used in real world sceaniors?
-How many meters apart 2 Wireless bridges at the same building have to be not to loose channel quality (although they will be used in different channels i'm sure if they are at the same location there will be some interferance)
-The buildings are about 2kms apart. I never tried 1310 series bridge before. Will 1310's can work 54mgbits speeds? Direct line of site and not much object on fresnel zone,some interferance can disturb us but thats why we used 21dbi directional dishes
The two bridge should be atleast 25 feet apart from each other, obviously on a different channel and different polarization(one set on horizontal and other on vertical). For a 2 kms distance the 21dBi gain directional antenna might be an overkill. I had the same issue when i tried these dishes for a 1kms distance. I ended up using the 1300s with integrated anteannas. Before going live you might want to do a test and make sure that the signal strength is not too strong. As per Cisco the signal strength should be anywhere between -40 to -50 for the best results.
54Mbps is not at all achievable. It is just on white papers. You would always have some interference and packet loss. Check with Cisco TAC and they will tell you what to expect practically.
I dont know much on how you can team up the two bridge setup to act as one big pipe. Cisco is good but I would rather find something that would give me the bandwidth I need without having to team up 2 seperate bridges. Check with Proxim they have some really good wireless bridge.
How did you arrive at the 25' number? Is this experience or is there a way to calculate or estimate this?
Cisco Fast Ether Channel (FEC) and IEEE 802.3ad are specific to 802.3 Ethernet and will not operate on top of 802.11 data-links. (But then, I never tried.)
I recommend you buy a WLAN solution engineered and advertised for the bandwidth and throughput you want.
I hope this helps. Thanks. /criss
Actually, according the BR1400 Deployment Guide, located here:
under "Stacking Bridges for More Throughput":
The bridge works in an integrated system environment where the attached Cisco switches or routers require aggregation protocols like FastEther Channel and Port Aggregation Protocol (PagP). FEC and PagP provide up to 100 Mbps of combined bandwidth.
During the installation, provide sufficient isolation for both redundant links by using different polarization or adding to the isolation, placing the two links more than 10 feet (3.048 meters) apart. RSSI voltage port verifies isolation.
According to the adjacent channel interference analysis, a system can tolerate a maximum input power of -34 dBm from an adjacent unit on a second-adjacent RF channel. With +24 dBm of output power, 58 dB of isolation between units is required to ensure satisfactory operation of each system. In other words, separation D between antennas must be adjusted to achieve this level of mutual coupling. Refer to Figure 39 to see the separation D depicted graphically and Figure 40 for an alignment graphic.
This configuration can also be applied to the BR1300, except that the BR1300 can only do .11b/g and the BR1400 only does .11a in the UNII-3 spectrum.
While I have not personally done this, I do know several Cisco SEs who have and reported successful, long-term customer solutions.
Well, it looks like someone did try putting two pairs of WLAN bridges between two Fast Ether Channel switches, and documented the result -- with caveats.
This "stacked bridge" configuration appears in a small section at the end of a much longer and otherwise excellent Cisco document on outdoors WLAN bridging between buildings. Overlooking the slip of defining "FEC" incorrectly as "Forward Error Correction" instead of "Fast Ether Channel", the language of this section is confused. The last sentence is typical: "The only reliable way to manage the bridges (ed. WLAN bridges) within PagP is to turn it off and route the traffic manually (ed. ?) by configuring the switches properly (sic)." We might be better advised to use a proper configuration in the first place and leave it at that.
IEEE Std 802.3-2002, Section Three, Clause 43 Link Aggregation, Subsection 43.1 Overview, reads: "This clause defines an optional Link Aggregation sublayer for use with CSMA/CD MACs. Link Aggregation allows one or more links to be aggregated together to form a Link Aggregation Group, such that a MAC Client can treat the Link Aggregation Group as if it were a single link. To this end, it specifies the establishment of DTE to DTE logical links, consisting of N parallel instances of full duplex point-to-point links operating at the same data rate." In this context those are full duplex Ethernet links only, with no intervening half duplex 802.11 bridges.
Cisco should nip this goof in the bud and apologize.
I hope this helps. Thanks. /criss
If it's in the documentation then it's been implemented and tested and is supported by TAC, otherwise it wouldn't be there.
Also, I submitted a correction to the FEC acronym error as I was reading the document. Should be updated quickly.
I don't doubt that what is described as "stacked bridges" has been done and passes traffic. I commend you for bringing the matter to our attention. I commend you for not having recommended anyone stack his own bridges.
What I mean is that Cisco goofed in allowing this configuration to be published in a way that appears to recommend it. To me "stacked bridges" looks like a junior network engineer experiment that escaped proper review.
As a comparison there are many duplex mismatches between Ethernet link partners that have passed traffic for years. And there are senior network engineers who insist on hard setting full duplex on their end of every Ethernet link even when the other end cannot be hard set to full duplex and is guaranteed to choose half duplex.
Inserting anything between two IEEE 802.3 Link Aggregation Control Protocol (LACP) aggregators other than some number of parallel full duplex Ethernet links is unwise as it both violates the IEEE standard and is unsupported by any proprietary extension to that standard.
I hope this helps. Thanks. /criss