If you're not gonna use the second card slot in the AP2000, we've found the AP600
works great (about 1/2 the price).
Also, Indiana University was in charge of networks at Internet2 this year in October.
At max usage, we had 300+ users on the wireless network. We used 3 indoor Vivatos
(one on each channel - 1, 6, 11 - pointed into the middle of the conference area) that
did GREAT with coverage, although there was some issues with hidden node that we
couldn't effectively quantify. (I'm happy to share details on that little experiment
- [EMAIL PROTECTED], 812-855-1784).
At one point we swapped over to a AP environment - we had 8 APs strategically placed
throughout the conference area. In square footage, I'd guess it was quite a bit
larger than two adjacent, 200 seat classrooms (two floors, too). These 8 Proxim APs,
configured in basic set-up, handled 300+ users, also. (Just for clarity, there were
800 participants, but DHCP logs showed less than half the users were concurrently
logged on).
Kirt Guinn
Indiana University
-----Original Message-----
From: 802.11 wireless issues listserv on behalf of Sean Che
Sent: Fri 1/9/2004 5:03 PM
To: [EMAIL PROTECTED]
Cc:
Subject: Re: [WIRELESS-LAN] Wireless Networking in Large Classrooms
John, Philippe, E.J. and all,
Thanks much for the advices!
I love this place!
a 3.5 dbi Omni-directional antenna 50 bucks
a pigtail for proxim radio card 20 bucks
a Proxim AP2000 b/g, 600 bucks
......
Your experience....and willing to share .... Priceless! :-)
Sean
Jonn Martell wrote:
> Our RF research group verified that you can go with a 4 channel
> assignment with 802.11b (DS) because the overlap is minimal. With
> 802.11g, this changes because of the way that OFDM work so you are stuck
> with 3 channels (testing pending to verify this).
>
> The IEEE uses 3 channels at its 802 conferences.
>
> Putting two radios so close together can be an issue because it create
> interference (even on non-overlapping channels). If the AP2000 still
> use client cards, you almost want to get some copper foil to wrap the AP
> 2000 client cards to reduce the leak of these cards (the internal
> antennas are supposed to be turned off when you plug external antennas
> but we found this was not the case). You can verify this with a
> spectrum analyzer - it's possible they fixed this. I'm not convinced
> that the AP2000 was really designed to have two radios in the same
> frequency range, it's really designed to have one in 2.4 and the other
> in 5.15-5.35 (but the marketing people will likely disagree on this).
> Test it for yourself if you want to be 100% sure.
>
> I would start with a single radio in 4 AP with the Proxim AP2000.
> (Actually, I would get 4 Cisco AP1200 (:)
>
> Disclaimer: The last time we tested the AP2000 was close to two years
> ago during our extensive RFP process so things might have changed and
> improved dramatically.
>
> As for AP placement, I don't think corners are the best place; divide
> the room into half or quarter (depending on the shape) and use
> directional antennas (as someone suggested) with the appropriate
> antennas and beam width to provide the coverage to each zone. It's hard
> to comment without seeing the shape of the hall. You won't be able to do
> too much about client-client and client-to-AP interference (as people
> indicated previously) but you can mitigate the AP-AP interference with
> proper RF planning and the proper antennas. Since a lot of the traffic
> is AP to client, it can work fairly well.
>
> Cisco has a clear advantage here because the RF performance is much
> better than Proxim (although Proxim has some great products :)
>
> Some links for good antennas and connectors
> http://www.superpass.com and http://www.hyperlinktech.com/
>
> Hope that helps.
>
> ... Jonn Martell, UBC Wireless, www.wireless.ubc.ca
>
> Sean Che wrote:
>
>> 800 people! Wow! I once went to Internet2/Joint Tech meeting and saw
>> similar sight, but that was about 400 people or so...
>> John, for that large lecture hall with 250 users, we plan to use Proxim
>> AP2000 with two radio interfaces. Four AP2000s are going to be
>> deployed, one for each corner so totally 8 radios. In such an open and
>> relatively small space, allocation of 8 channels seems not trivial,
>> e.g. not possible for 1-6-11 scheme Most of the channels are going to
>> be used . Some of the overlapping radios ( basically they are all
>> overlapping :-P ) have to use adjacent channels. Any
>> suggestion/experience about this?
>>
>> Thanks,
>> Sean
>>
>> Jonn Martell wrote:
>>
>>> I used to be very worried about high density until I started to attend
>>> the IEEE meetings a few years ago where there is close to 800 engineers
>>> with laptops downloading PDFs, PPTs and DOCs. Quite the sight! I wish
>>> there was a way to take pictures but these aren't allowed at IEEE
>>> meetings. Worth the trip to one of their conference as an observer if
>>> you want to increase your comfort level on high density deployments.
>>>
>>> Every wireless engineer has a laptop and they are all in the same
>>> ballroom at the beginning and end of the conference. During the
>>> conference, all the attendees are in close proximity as the large
>>> conference hall gets broken up into a dozen smaller large meeting
>>> rooms.
>>>
>>> I'm not convinced that tuning the radios below the power of most
>>> clients
>>> is a good idea and our RF research group has found that power
>>> control in
>>> its current state is really inadequate (as a result, we aren't focusing
>>> on power tuning in our deployment).
>>>
>>> To do load balancing, the trick I think at this point is to make sure
>>> that you turn off support for the lower speeds to force roaming to the
>>> other stronger APs. There is no standards-base way of doing load
>>> balancing.
>>>
>>> What the IEEE is doing with IEEE 802.11k is an attempt to provide a
>>> standards-based resource management information so that radios can help
>>> tune down the power of clients (as it's done in the cell phone
>>> industry)
>>> so that clients don't keep blasting away if they don't have to. So this
>>> problem is getting fixed because the market needs it. I'm not too sure
>>> if the problem is going to be fully fixed with 802.11k but Cisco, with
>>> its "Cisco Compatible" CCX program, is doing the same today. They are
>>> just ahead of the slower moving standards bodies but now have several
>>> vendors supporting CCX (this list was empty last year at this time).
>>>
http://www.cisco.com/en/US/partners/pr46/pr147/partners_pgm_partners_0900aecd800a7907.html
>>>
>>>
>>>
>>>
>>> Until this is widely available, directional antennas at the APs for
>>> these special circumstances makes a lot of sense.
>>>
>>> For large theaters, we deployed a single AP for now but we have
>>> three AP
>>> drops (each AP drop has 2 cable/circuits) so we can scale to 6 APs
>>> if we
>>> need to.
>>>
>>> I predict the ultimate answer for high density in large rooms will be
>>> the next generation of 802.11a possibly combined with standards-based
>>> client radio management. In the 5 GHz WLAN spectrum there is 200
>>> MHz of
>>> available spectrum versus just 83 MHz in 2.4 GHz range. IEEE 802.11a is
>>> just not there today...
>>>
>>> ... Jonn Martell, UBC Wireless, www.wireless.ubc.ca
>>>
>>> Sean Che wrote:
>>>
>>>> High density is a big challenge to wireless deployment. We are
>>>> currently
>>>> facing the same issue. In one of our wireless projects, we were told
>>>> that there might be up to 250 simultaneous users ( Even worse: Did I
>>>> mention they are all Pocket PCs with wireless cards? ) in one large
>>>> lecture hall for class. In this kind of "noise" crowded environment,
>>>> not only the APs will interfere with each other, the clients radio
>>>> cards
>>>> will also join the choral society.. What a nightmare!
>>>> We are thinking of using directional antennas to help distributing
>>>> the
>>>> clients evenly; tuning the transmitting power to minimum. The problem
>>>> is we couldn't really get a feeling how it works before we really
>>>> install it and those 250 students really start using it ( and maybe
>>>> complain about it. )
>>>>
>>>> Sean
>>>>
>>>> Arnold Hassen wrote:
>>>>
>>>>> We are designing two new 200 seat classrooms that will be adjacent to
>>>>> one another. Discussion is focussing on whether we should
>>>>> hardwire or
>>>>> go wireless.
>>>>> Functionally we must be capable of simultaneous networking which
>>>>> means
>>>>> 400+ simultaneous links.
>>>>> Is this doable with wireless?
>>>>> Thanks for any help
>>>>> Arnie Hassen
>>>>> West Virginia School of Osteopathic Medicine
>>>>> ********** Participation and subscription information for this
>>>>> EDUCAUSE Constituent Group discussion list can be found at
>>>>> http://www.educause.edu/cg/.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> -------------------------------------
>>>> Sean Che
>>>> Network Engineer
>>>> Network Services
>>>> Wayne State University
>>>> Voice: (313)577-1922
>>>> Pager: (313)990-5403
>>>> Email: [EMAIL PROTECTED]
>>>> -------------------------------------
>>>>
>>>>
>>>> **********
>>>> Participation and subscription information for this EDUCAUSE
>>>> Constituent
>>>> Group discussion list can be found at http://www.educause.edu/cg/.
>>>>
>>>
>>> **********
>>> Participation and subscription information for this EDUCAUSE
>>> Constituent Group discussion list can be found at
>>> http://www.educause.edu/cg/.
>>
>>
>>
>>
>> --
>>
>> -------------------------------------
>> Sean Che
>> Network Engineer
>> Network Services
>> Wayne State University
>> Voice: (313)577-1922
>> Pager: (313)990-5403
>> Email: [EMAIL PROTECTED]
>> -------------------------------------
>>
>> **********
>> Participation and subscription information for this EDUCAUSE Constituent
>> Group discussion list can be found at http://www.educause.edu/cg/.
>
>
> **********
> Participation and subscription information for this EDUCAUSE
> Constituent Group discussion list can be found at
> http://www.educause.edu/cg/.
--
-------------------------------------
Sean Che
Network Engineer
Network Services
Wayne State University
Voice: (313)577-1922
Pager: (313)990-5403
Email: [EMAIL PROTECTED]
-------------------------------------
**********
Participation and subscription information for this EDUCAUSE Constituent Group
discussion list can be found at http://www.educause.edu/cg/.
