Slashdot Mirror
WiFi Offloading is Skyrocketing
dkatana writes: WiFi Offloading is skyrocketing. This is the conclusion of a new report from Juniper Research, which points out that the amount of smartphone and tablet data traffic on WiFi networks will will increase to more than 115,000 petabytes by 2019, compared to under 30,000 petabytes this year, representing almost a four-fold increase. Most of this data is offloaded to consumer's WiFi by the carriers, offering the possibility to share your home internet connection in exchange for "free" hotspots. But this article on InformationWeek Network Computing also warns that "The capacity of the 2.4GHz band is reaching its limit. [...] the growing number of WiFi devices using unlicensed bands is seriously affecting network efficiency. Capacity is compromised by the number of simultaneously active devices, with transmission speeds dropping as much as 20% of the nominal value. With the number of IoT and M2M applications using WiFi continuously rising, that could become a serious problem soon."
The real reason for the cell companies to "offload" data is to ease the load on their networks. OK, I understand that ... and I realize that it could save me money. If I'm in a hotspot, why not use that instead of eating my limited data plan?
But honestly, Verizon has almost gotten ridiculous with it. Little beg screens ("are you SURE you don't want to connect to wireless?" -- it was a happy day when I figured out how to kill that one), refusing to open Web pages if I'm just beyond the range of a known hotspot, and worse.
Verizon is VERY aggressive about offloading.
Given how much it costs to build a new tower site nowadays, I can understand, but don't be fooled: the benefit of offloading is primarily for the cell carrier, and NOT for you. :)
Cogito, igitur comedam pizza.
coming from your IP
You know the public hotspot traffic is segregated to a separate IP addresses, right?
At least that's how Comcast does it. Can't imagine there rest aren't also doing the perfectly obvious.
There are entirely legitimate reasons to object to this stuff, but being held liable for public hotspot traffic due to conflated IP addresses isn't one of them.
Maw! Fire up the karma burner!
One of the Part 15 rules is about not interfering, and clearly that's not working out so well.
I don't believe the FCC cares about Part 15 devices interfering with other Part 15 devices. From the rules:
CFR 47 Part 15.5
(b) Operation of an intentional, unintentional, or incidental radiator is subject to the conditions that no harmful interference is caused and that interference must be accepted that may be caused by the operation of an authorized radio station, by another intentional or unintentional radiator, by industrial, scientific and medical (ISM) equipment, or by an incidental radiator.
(boldface is added for emphasis)
In other words, a Part 15 device that interferes with another Part 15 device is cool, since Part 15 devices are offered no protection from interference, intentional or unintentional.
The law is written so that Part 15 devices are not allowed to interfere with licensed devices.
I run the campus and dorm WiFi system at the university that i work at. We have many 250 room(500-750 people) residence halls that use around 30 3x3 802.11N dual band (2.4 and 5.8) access points each (we have a total of ~270 APs in 9 halls). The 2.4Ghz band is perfectly usable ALL of the time(at least 1.5Mbps usable data throughput at all times). Our students usually have around 2000-2500 devices on each band (we allow EVERYTHING except bit torrent) with game systems/appleTV/anything....we do not disallow any device as long as it doesn't impact other people.
Now how do we keep 2.4 usable in an environment like that? We manage the crap out of it. First we only allow channels 1,6,and 11 to be used. This keeps anyone from stepping on two channels at once. We also (and this one helped a lot!) turned the N protocol bandwidth width to 20Mhz instead of 40Mhz. At 20Mhz, you have 3 completely separate channels. When the width is 40Mhz you basically chew up 2 open channels at once.....so all channels are always walking over each other. Although this does decrease raw throughput of a client, it almost doubles the amount of usable connections per radio and helps a lot with the further away users from the AP. We also do 5 minute power and channel tuning where the system keeps all that balanced as usage differs. We also do not allow anything slower then 11Mbs to connect at all. This has a two fold benefit. One being that it makes sure the client/AP stays with the closest one to it and two...and this is the HUGE biggie....it increases the efficiency of the 802.11 time slice distribution. 5.5Mbps requires 10ms per client in radio chat time that the client gets regardless of how much data it is sending or receiving. 2 and 1Mbps requires 20ms! The old 802.11b is horribly inefficient and actually causes less clients to be able to connect to a given AP. Turing of everything but 11Mbps increased AP/client concentrations around 40%. Another thing we do is NOT impose is per connection speed limits. We found this reduces amount of usable connections per AP(about 10%) and slows down everybody for no benefit 802.11 is good at balancing throughput between all clients already. We also force transfer anyone off the 2.4 band that is 5.8 capable. The last thing we do is have alarms for when the system shows very high levels of interference on the 2.4 band. Sometimes this is a bad cordless telephone or something but 90% of the time......it is a stupid microwave that went bad. In dense university living areas, they are EVERYWHERE!! Especially since the students buy the cheapest ones they can find. One bad microwave will kill 2.4 in an entire area.
Now...all that being said...when people tell me that 2.4 is crowded and slow because it is unlicensed, I tell them no....it's just managed very, very badly. I have been wishing that one day in the future...hopefully soon as i have requested it during the public comment phase of the last few 802.11 standards...that good spectrum management would be added to the WIFI standards to help with this. The best they do now is look around and try not to be on the same channel. I wish there could be some extremely low level protocol where all access points as part of spec, discuss the rf environment and attempt to keep it sane instead of the apartment fun of 100's of APs all blasting at MAX power to try and get it's voice heard over everyone else's voice. Last time i went to a friends apartment, he was complaining his wifi was slow. When i looked at the rf information.....the noise floor was freaking -62dbm. I told him it will never work right in that environment.