It's one seamless network in the "Roaming" configuration.
How are the signals on the mobile devices when roaming?
You signal should never drop unless you are out of range of your entire network...
As long as you have a good signal, then you are fine.
I'm sure you read this Apple article on Roaming....
The signals are acceptable. However, the issue is more of the device "switching" from one access point to another. For example, if my iPhone4 is fairly close to one of the APX devices in my home, I might have signal ranging from -15 to -30 or so. If I then move to another room with another APX, my iPhone will stay associated with the first AXP, but signal will drop to -50 to -60 or so, and won't automatically jump to the APX in the new location. However, if I toggle the wireless mode on the phone off and back on, the iPhone will then associate with the closer APX, and will have better signal strength of the -15 to -30 range.
It would make more sense to me if this switching between AXP's occurred automatically, rather than forcing this manually by toggling the wireless off and on.
Yes, I did read the Apple article on roaming (and many others from different sources, as well).
As I mentioned, it is one seamless network.
It is receiving a signal from one network.
If your signals are good then there is nothing wrong with your network.
In this configuration, it doesn't have to switch to different units to work.
No need to connect to different boxes as it connects to the same wireless signal basically.
While it is true that I have created a single, roaming network, with a single SSID, the performance of the device (i.e., bandwidth), as measured by data throughput, will vary with distance to the various wireless access points (in this case, APX devices). My device will be connected to the network via an association with one of the APX's at any point in time. Therefore, if I my iPhone is associated with an APX that I am physically close to, I will get optimum bandwidth (as measured using one of the many apps that assay this, such as SpeedTest). However, if I move to the other side of the house, and am still associated with the original APX device, I will still be connected to the same network, but will get a definite performance drop. On the other hand, if my iPhone associates with another, closer AXP on my network, I will maintain optimum performance, rather than degraded performance that occurs by being associated with an AXP that is further away.
The reason that well-designed wireless networks, such as those in businesses, have many wireless access points, is so that one can remain connected to the network from various points in space throughout the enterprise. Typically, the device will "connect" with the access point with the strongest signal (usually the one closest to the user), although this process is basically seamless, and the user only "sees" that he/she is connected to the SSID identifying that wireless network.
Therefore, while I agree with you that I am connected to "one seamless network", behind the scenes I should be jumping off and on to the various access points that comprise the network based on signal strength. It is this function that seems to be somewhat compromised in my case.
"It is this function that seems to be somewhat compromised in my case."
No it isn't.....
Why would there be a need to do that?
It's all one network...
It is true that they will connect to the closet device when be turned on...
However, it doesn't matter.....
Try this in reverse....
Turn your device off and go to your furthest access point...
Turn the device on... It is probably is now connected to that closet device...
Now walk into the room where you main device is.... It probably will still be connected to the original device it was connected to and NOT change....
Why would it need to?
It still has a strong signal from the one and only network you have...
I don't think I can tell you anymore.
You have likely noticed that it is the mobile devices that tend to "hang on" to the original wireless access point that they associate with.
Some of this is due to their operating system, which is less sophisticated than the operating system on a laptop computer.
I imagine that it would be a pretty good bet that the Apple engineers set the thresholds differently on mobile devices compared to laptops. Unfortunately....if there might be a way to adjust these levels....Apple is not telling.
Yes, Bob. That is true. My wireless laptops tend to reassociate with closer access points more readily and intuitively than the Apple mobile devices.
With a bit more experimenting, I have found that there appears to be a threshold signal level, somewhere around -65 to -75 dB, where the mobile device will "let go" of the original access point, and "connect" to the best (i.e., closest) one. Too bad this threshold isn't configurable, though, since my testing does confirm some loss of throughput between connections at, say -55 vs -25 dB.
I need help, I work for a school that just went all MAC, we have cisco access points through out the school...in excess of 200 of them....There are many areas that overlap which fine, SSID all the same etc....to provide what should be a seamless roaming environment. However we have found that many...almost all actually will not operate "seamlessly" they will drop one access point and then connect to another, this takes anywhere from 30 seconds to a few minutes.
Our windows machines don't do this. Any thoughts of why our Macs, Ipads, Iphones are doing this and not as they should???
I agree that the lack of roaming is worrisome. It does not work as expected and shoudl work "better" but realize that here may be implications to it "working better":
When you have several Airport base stations of various types all using the same SSID and settings, one would expect that devices "roam" and disconnect from one unit and reconnect to a new unit whenever it is more attractive than the existing one the client is connected to....
The new one should be measurably more attractive or you could be flip-flopping all day long.
The device should only reconnect to a new one if no data is being transmitted or received (this is a biggie) or your streaming video, email send etc will be broken mid-stream...
So its likely the change will only happen when the device is "sleeping" or when the existing signal gets so bad its likely to cause a disruption anyway.
I say attractiveness since its not all about signal strength:
- Better in terms of signal strength... (Stronger signal/lower noise is better... but....)
- Better in terms of the bandwidth of the signal (5GHz band is better than 2.4... depending on distance....)
- Better in terms of upstream connectivity (100Mbps Airport vs 1000Mbps Airport Extreme uplinks... I prefer a 1000Mbit unit if I can transmit at >100Mb/s, no?)
- Better in terms of having fewer clients connected/ being less busy.
So, to figure out if a client should roam, the client needs to, in a perfect world:
- Wait till you’re not doing anything....
- Tune the antenna to the new potential Airport's frequency band.
- Ask the new Airport how many users are on it, how fast its connected and how busy it is etc.
- Do same for all other airports....
- Take the existing Airport's characteristics (of Bandwidth, Business, Signal strength) and compare that to all new candidates to decide if there’s a better one.
- Switch! While realizing that if they roam, connectivity (the video playing problem) is interrupted, while the device gets a potentially new IP address etc (Whose to know if that other Airport is on the same network?)
So, as you can see its quite complex... but yes, it "should work better". Enterprise systems for a couple of grand do this great.... but not for $100 or $200 a pop.
iOS and to a degree Macs, seem not to actively look at this data to make an informed choice.
I'm experiencing exactly the same problem as you Stuart. Having 4 different Airport Expresses in the house , located pretty far away from each other. Not only is my iPhone and iPad having this problem, but my Android smartphone is having it too. To a worse degree, it does not switch to another access point, even when the level is so low, that reception is almost non existent.