we are looking to roll out wireless for voip and handhelds around our environment however senior management are leery about the cost of a site survey and so they want to do it in pieces, paying as they go.
Does anyone out there have any documented evidence that this is a bad idea. Common sense says you should do the whole building at once (it is a 9 storey hospital) but I need evidence to support my proposal of not breaking the survey up. My thought is that there will be some problem stitching the surveys together if you don't have it all done at once and that there may be dead areas around the perimiter of the survey areas.
Please help out if you can... if you wish, email me directly at skay AT nygh.on.ca
What cost? Site surveys are done based on the probable location of access points. Go down to the store, buy an access point (or wireless router), find a wireless laptop, and do some plotting. Move the access to all the probable locations and record the signal strength at each location using Nestumbler or Kismet.
I've done a few of these in office buildings and got a surprise. The timer burner is not in doing the survey. It's in arguing over the probable number location of the wireless access points. You can't just plaster the place with a zillion wireless access points because they interfere with each other. You also can't just run the whole building on one wireless access point. Obviously, there's an optimum number and that number is usually not obvious. If in doubt, pick 3 access points per floor (on channels 1, 6, 11) and you've hit maximum.
Unless I'm missing something here, you're looking for documentation for the wrong way to do something or documentation of a failed installation. It's possible to find those, but I don't know of any. Actually, I do know of some failures, but they weren't because the site survey was performed in an odd manner. It was because no site survey was performed in the first place.
Can I render my opinion in the matter? My experience with hospital construction is that each floor can be considered a separate RF environment. There is so much concrete, steel, foil backed insulation, and shielding in the walls, that nothing penetrates the walls. Same with the floors and ceilings. The biggest headache I had in planning a hospital wireless system was getting sufficient coverage through all these barriers to RF. For example, we had a waiting room and lobby with a series of office radiating from the center. I installed an access point near the middle of the lobby on the theory that it would also cover the offices. It did, but only if the doors were open. I could be 5 meters away, but if the door was closed, wireless was a lost cause. The floor layout also included a substantial number of zig-zag corridors. The idea was to keep the noise level down by not creating a tunnel. That worked fine but blocked RF going down the corridors. I setup one wing of the hospital to be "illuminated" from an access point located OUTSIDE of the hospital on a nearby roof. That worked better than trying to light up the rooms from inside. Like I said, location of access points is far more difficult than actually performing the site survey. Of course, it's useless to perform a site survey without first knowing where the access points and users are going to be located.
True. However, there are going to be dead areas no matter what you do. The construction of the hospital largely creates this condition. It's not like corporate offices, which tend to be more like a partition infested warehouse, that can be illuminated from above the partitions. It's more like a labyrinth, dungeon, or bomb shelter.
If dead zones are as bad as I predict, you might want to look into "leaky coax" cable.
These can be snaked through the overhead utility areas. They really only offer coverage for short distances (about 3 meters) but are perfect for snaking corridors and labyrinths.
I'm not sure what you're thinking of in the way of VoIP Wi-Fi handsets, but methinks you should try using one for a short test before committing the hospital to what I consider to be an R&D project. Wi-Fi VoIP is very new technology, with lots of potential for surprises. Try it before you buy it.
Wanna pay for my time and advice? I answer questions in public newsgroups and mailing lists for free. I charge for personal consulting via email or phone. Unless proprietary info is required, answers should be posted where everyone can learn from both the questions and answers. Otherwise, it's consulting.
Thinking of rolling VoIP over wireless for your hospital, you will save a lot of communication money between staff members, it worth doing it, but doing in the right way.
I come to a attend half a day ARUBA seminar and I think those guys really have an exciting story to tell, they have a distinguished solution for your case with very high QoS for these real time application with very high security.
check their site for more details or even call them and listen to them ( you are not going to lose any thing)
I've had the opposite experience with concrete floors and ceilings in multi-story builings - too much "bleed-through" between floors!
Only having three non-interfering channels available (ie 1, 6 and 11) is usually okay for a single floor, but it's tough when the signals from the floors above and below are still strong enough for devices to connect to.
Yes, that can happen. However, if you do some direction finding, you'll probably find that the signal is NOT coming through the concrete floors or walls. The RF attentuation of cast concrete at
2.4GHz is fairly high:
What I found is that the signal goes out the windows horizontally, bounces off a nearby building, and re-enters the building through windows on the same or different floors.
It's particularly bad in multi tower office buildings with glass curtain walls. Aluminized mylar energy saving coatings on the glass help block some pickup through the windows. Reflections are not as bad as direct 2.4GHz interference through the windows from a nearby building. At one local office building complex, I counted over 100+ SSID's. There's nothing that can be done with 3 non-overlapping channels to prevent interference in such an RF crowded environment.
I've also used alternative methods of 2.4GHz distribution in a hospital. See item #4.
I wouldn't try this today because there the signals were fairly weak and there's far more potential for interference today.
Perhaps a different type of site survey might be useful. See:
This cheap spectrum analyzer will display any sources of 2.4GHz RF junk, no matter what modulation type. It's not particularly sensitive so if it sees any interference, it will probably become a problem. Such nightmares such as a frequency hopping (Uniden) cordless phone:
will show up on the Wi-Spy spectrum analyzer, but not on Netstumbler or Kismet. Note that the phone clobbers all the 2.4GHz channels.