Cool-Tether Links Phones' Bandwidth To Make High-Speed Hotspots 102
Barence writes "Microsoft Research has found a novel way of beating the deplorably slow speeds of mobile broadband, by combining several phones together to make one high-speed hotspot. Dubbed Cool-Tether, the system harnesses the mobile data connection of multiple mobile handsets to build an on-the-fly Wi-Fi hotspot. 'To address the challenges of energy efficiency, Cool-Tether carefully optimises the energy drain of the WAN (GPRS/EDGE/3G) and Wi-Fi radios on smartphones,' Microsoft's research paper claims. 'We prototype Cool-Tether on smartphones and, experimentally, demonstrate savings in energy consumption between 38%-71% compared to prior energy-agnostic solutions.'"
like BitTorrent (Score:5, Insightful)
a novel way of beating the deplorably slow speeds of mobile broadband, by combining several phones together to make one high-speed hotspot.
Mobile operators will just love this! Considering the cell towers can be a bit slow already and especially so when many people are using them for internet, this will not magically provide better speed off it. But it lets users abuse the network same way that BitTorrent does - hammer the network so much that you get more while others suffer.
While operators already have unlimited 3G for cheap (not in USA, so they actually are unlimited), the only way slow speeds of mobile broadband is going to improve is to push for new technologies and make the operators improve their network. But not that 3G's 5Mbps would be that slow anyway.
Re:like BitTorrent (Score:5, Interesting)
I was actually thinking of something like this yesterday. With the rapid increase in Wifi + Internet enabled phones and devices, it could be possible to actually have an entirely distributed network just by linking together devices in range.
Perhaps that's where we should build the Internet 2, now governments around the world are doing everything they can to control the first one. :)
Re: (Score:3, Insightful)
With the rapid increase in Wifi + Internet enabled phones and devices, it could be possible to actually have an entirely distributed network just by linking together devices in range.
And just imagine the legal complexities if someone actually ran a torrent over it, with unapproved content...
Re: (Score:2)
Do I win?
mesh networks (Score:2)
Re: (Score:3, Informative)
Re: (Score:3, Funny)
But it lets users abuse the network
Bittorrent?
I like to abuse my network by complaining how slow it is to responding to my requests for pictures of sandwiches and how much space its old equipment takes up. I always threaten to keep it off the surge protector or knock it off the shelf so I can get a nice new slim model with all the bells and whistles.
Re: (Score:2, Informative)
Specifically, the issue is that HSDPA only gives about 3Mbps per tower, and no mesh wi-fi network will get around that because each phone will be using the same over-subscribed tower.
Re: (Score:3, Interesting)
Re: (Score:2)
Re: (Score:3, Informative)
I'm pretty sure you mean per channel. Multiple devices can use different frequencies from the same tower.
Re:like BitTorrent (Score:5, Insightful)
Using bandwidth that you have PAID for is not abuse. A company overselling their capacity or promising more bandwidth than they provide is fraud however.
Re: (Score:2, Interesting)
Re: (Score:2)
You guys missed the redefinition of "unlimited" Internet... now "unlimited" means you can use it whenever you want and be connected all the time (unlimited connection!), for whatever low-bandwidth tasks the ISP approves of. To reserve bandwidth for their VOIP or IPTV or whatever products/services.
Re: (Score:2, Interesting)
Re: (Score:2)
You're correct that ISPs (like most businesses) are greedy. However, you're completely mistaken about the scale involved.
You can currently buy 10mbps cable service in most markets for under $100/mo. A real, dedicated 10mbps link with guaranteed bandwidth costs about $1,000/mo. Even that just gets you priority, not a full guarantee, when it comes to backhaul traffic.
Let's do the math. A mid-sized metropolitan area might have a half million households. Say 10% of them had internet access - that's 50,000
Re: (Score:2)
Half true. The US is about as big as all of Europe. With some of the lowest population density of any civilized nation, it takes a lot more cabling, and a LOT more infrastructure.
Re: (Score:2)
Meaningless...
We should be able to compare pricing between, say, Oxford and Austin. Or NYC and Paris. Rural pricing would differ due to density, urban pricing for similar densities shouldn't vary for that reason (it may for other reasons, of course). And the rural argument falls short considering the billions the telcos have been given by the government (through additional mandatory charges to their customers) to cover their cost in upgrading the rural network.
Re: (Score:2)
So you can connect at 100mb to the other side of NYC. To get to a server in California, you still need nationwide infrastructure.
Re: (Score:2)
Re: (Score:2)
Depends on how wide your definition of rural is. The real shortcoming we have in the US is small cities in the 5,000 to 100,000 population range. And they most don't have links to them of any grade one would call "backbone."
Re: (Score:2)
...small cities in the 5,000 to 100,000 population range. And they most don't have links to them of any grade one would call "backbone."
True enough, as a city dweller I often forget that. I suppose I was commenting on getting from NYC (which definitely has a tier one backbone going into it) to a city in CA sounds like it should be going over a major pipe that can handle the load, at least until it gets to CA and has to possibly route into a tiny town, which goes back to my comment of the problem being local POPs with low capacity. Granted you are entirely correct that some low population ranged places are likely not going to have nearby lin
Re: (Score:2)
Yeah, more places in this country have been visited by Google's Street View vans than have ever had access to high speed Internet.
Re: (Score:3, Insightful)
Or, they could do what any sane person would do, and realize that at any given moment only a tiny fraction of their users are using ANY bandwidth
Exactly. And that's precisely what makes the whole not-enough-bandwidth problem so ridiculous. No one's asking them to provide the total theoretical amount of bandwidth that they're selling. But they're overselling by so much that they can't even cover what their customers ARE trying to use, let alone what they're actually selling. It's basically the equivalent of an airline selling 500 tickets for a 120-passenger flight. Not to mention the fact that (in the US) they have already been paid by the government
Re: (Score:2)
I wasn't really talking about using the bandwidth, just the way how BitTorrent establish it and what effects it has on the network. They both "spam" the network with as many connections as possible to maximize speed, but that comes with cost of the others. On local level, if your torrent client is using 1000 connections at a time, your browser that is using 1-5 connections is going to suffer.
This thing is doing basically just the same, only that the ones "spamming" the phone tower with connections are those
Re: (Score:1)
Re: (Score:1)
QoS NEVER works.
Re: (Score:2)
QoS NEVER works.
Not true. It works until the first time somebody decides to abuse it.
Oh, wait. That's what you said, isn't it. Never mind...
You get what you pay for. (Score:2)
Using bandwidth that you have PAID for is not abuse. A company overselling their capacity or promising more bandwidth than they provide is fraud however.
You might want to read your contract and TOS.
If you are paying the mass market price for broadband you are paying for speeds "up to" some limit.
When and as available.
Western Union - a century or so back - printed a disclaimer on the top of every ordinary telegraph form that promised nothing more than a good faith attempt at prompt and accurate delivery.
The
Re: (Score:1)
The bit-torrent argument is slightly inappropriate due to the fact I do not see constant, all-day file-sharing becoming common place on mobile phones in the near future. (primarily due to battery issues) This is for downloading an attachment, or more likely, many separate images to load a website faster. I do not expect many people will use this technology to download a blu-ray movie from mininova with an down/up ratio of 1:1.
It may enc
Diffrent phones : Different operators (Score:2)
Well, nothing prevents you from using several phone each using *a different* operator.
Thus you're not eating up more "user slots" than the average user, but spreading your bandwidth across several towers of several operators.
And the bittorrent problem is more linked to shitty service providers who attempts to oversell more bandwidth that they actually have and then come back crying when the users start attempting to use their connections as advertised.
Re: (Score:2)
A better comparison would be to "download accelerators". You know, the ones which boost download speeds by opening 30 connections to the exact same server, crowding out other users...
Bittorrent isn't designed to crowd out small users off of the net in order to get its speed, it's just an unfortunate side-effect of how some network hardware works. Download accelerators on the other hand, get their speed exactly the same way as this device.
What mobile company would support this? (Score:4, Insightful)
Re: (Score:1)
They can't hold off supporting it much longer. As more providers distribute phones capable of supporting features like this, they are all brought closer and closer to the line. One of them will cross it first, and then the rest will follow. Tethering will be a ubiquitous feature in another year or two, even in the US.
No, they don't like it. Yes, they will do something about it if/when they perceive it to be abused. No, it won't matter in the long run.
Re:What mobile company would support this? (Score:4, Insightful)
Re: (Score:2)
How would they stop you? Androids, iPhones (once jailbroken), Windows Mobile, Symbian, Maemo, etc. all let you run your own apps. How hard would it be to write one (and port it to all the above platforms) that implements this with a common/standardized protocol? Just combine the program with a browser and you don't even need to worry about messing around with the OS
Meme redux (Score:2, Funny)
Can't wait for the pwnage (Score:2)
suggestion... (Score:1)
Re: (Score:2)
won't that just cause people to want to lick their iCandy?
But, but....... (Score:5, Insightful)
Re: (Score:2)
Logical progression != innovation
Re: (Score:2, Interesting)
Re: (Score:2)
You Mods do realize that the differences in Office 2010 and Office 2k8 are interface differences?
Nothing is more likely to have a direct impact on the productivity of the office worker than the UI.
You do realize the differences in Vista and Windows 7 are mere "bug fixes", much like Win 95 and Win98 were.
These "bug fixes" have been enough to give Win 7 5% of the global market one month after its official release.
Windows 7 Breaks 5% in Daily Tracking - Mac Share Drops .15% in November [hitslink.com]
Re: (Score:1)
Re: (Score:2)
Re: (Score:1)
I thought about moderating this poster up, but decided that there's a fair chance I'll participate, so I'll back up this.
I do research in programming languages (almost more "program analysis" at this point). The two top conferences in our area are POPL (Principles of Programming Languages) and PLDI (Programming Language Design and Implementation). At least in my area, MSR (Microsoft Research) publishes at least on par with a top-tier research university, and judging by the program [psu.edu] for POPL 2010, even more s
Re: (Score:2)
They could if they weren't so greedy. I actually respect the fact that MS sponsors pure research, more or less. But they should look to GOOG to learn how to capitalize the results properly (instead of being douchebags).
No big innovation (Score:2)
Cydia (Score:2)
so when will they release this for Cydia?
...and the news is power saving? (Score:2)
The system is most likely to be harnessed in developing nations such as India, where mobile internet is far more prevalent than fixed-line access.
So, the system is aimed at applications where GPRS/EDGE/3G speeds are not sufficient but there is no access to power lines, and there are several phones to mitigate the speed problem? Like, I don't know, team of computer pirates torrenting while on the move to be hard to locate? Or live TV broadcasting?
I mean, usually if you have several smartphones at hand, and
AKA JoikuBoost (Score:3, Informative)
Way to innovate MS! JoikuBoost [joiku.com]: "JoikuBoost joins multiple 3G connections from mobile phones and operator networks into one larger unified and shared bit pipe, accessible over WiFi from e.g. laptops."
Who wants to bet they'll get the patent anyway ?
Re:AKA JoikuBoost (Score:4, Insightful)
Bill
Re: (Score:2)
Except for the fact that JoikuBoost doesn't manage the power utilization like Cool-Tether does. Also, this is actually a follow on to their previous invention called "Combine" which was announced in 2007. JoikuBoost appears to be a very recent (October 1, 2009) product.
Multipath IO is hardly what I'd call an invention these days (or even in 2007) but yeah that certainly puts them ahead of Joiku. Still you know what they say, "Real artists ship." ;-)
The energy savings do sound pretty spectacular though.
Re: (Score:1)
Alpine! (Score:1)
Deplorably slow? (Score:2)
Deplorably slow? HSDPA is already faster than most people's broadband. Seems like a solution looking for a problem to me...
Re: (Score:2)
Re: (Score:2)
I would say 5 mbit is faster than what most people have in many parts of the world. Especially in areas where ISPs don't oversell their bandwidth. Here in Denmark, most users choose a 2 or 4 mbit plan as it is cheaper and fast enough for them. It is also quite common to see HSDPA over 10 mbit (again, since the providers tend not to oversell much).
Hot-Tether (Score:2)
out of the box on Linux (Score:2, Informative)
You get this kind of thing out of the box on Linux: just plug in multiple phones and configure multiple internet connections; you get load balancing, on-demand dialing, and all that for free. Linux got this support years ago for dial-up modems, but mobiles phones look like dial-up modems to Linux anyway. It's not usually done with cell phones because it's expensive (that's why there's no simple UI for configuring it), but it's well documented and pretty easy to set up.
(Of course, with Windows and WinMo, i
Re:out of the box on Linux (Score:5, Insightful)
that's why there's no simple UI for configuring it
Well, for many people, they cannot do anything on a computer with out the "simple UI". So bringing something that a very small population knows how to do on a OS that most have not heard of to the general population might be something worth doing.
Re: (Score:1, Interesting)
You lost me at "plug in."
Good job (Score:5, Insightful)
Microsoft Research has found a novel way of beating the deplorably slow speeds of mobile broadband...
Good job, research division. Now reluctantly hand it over to marketing which will:
- Tie it to Windows Mobile
- Cripple it to only work with Hotmail and Bing
- Junk it up with "partner channels"
- Drag out deployment long enough for Apple to be able to field something smaller, cooler and 5x more expensive six months ahead
Re: (Score:2)
- Google releases an unrestricted similar product that's twice as capable.
- ?????
- Profit.
Don't you need seperate gateways for this? (Score:3, Interesting)
The other way to get around this is to have 2 routers working for you doing basically the same thing, but the speedup is only between those two routers. To get faster internet speeds I'm pretty sure separate gateways are needed. Do they get around this ?
http://lartc.org/lartc.html#LARTC.LOADSHARE
Re: (Score:3, Interesting)
I've considered similar for when traveling by train. Not necessarily multiple mobile phone connections, but at least one phone and a connection to the train's similarly reliable and very crowded wireless (the train wireless is sometimes noticeably more lethargic than a GPRS link).
The thought is a simple UDP relay/tunnel that can load balance packets over multiple connections (I have a little server out there that would act as the other end-point) and run OpenVPN over that channel for everything else. That w
Power utilization estimates sound like B.S. (Score:2)
savings in energy consumption between 38%-71% compared to prior energy-agnostic solutions
This strikes me as hubris (at least a little). While TFA talks about throttling down power usage, we're talking about doing this with multiple components being replicated (CPUs, maybe WiFi receivers, connectivity between phones, etc.).
MS is a Hog (Score:1)
Microsoft Research (Score:1, Troll)
The same brilliant minds that brought us Microsoft Songsmith.
Re: (Score:2)
So I get modded troll for mentioning another Microsoft Research product? Ok, so I forgot the link: Microsoft Songsmith [microsoft.com]
It's a program that plays accompanying chords to match up with improvised recorded singing. The video ad they created is even more painful of a concept.
Typical Microsoft academic pub- why? (Score:1)
As usual, they choose to address things readers will find interesting and leave out important details. Here's a few pseudo-equations for you:
PowerRequired(802.11) < Power(3G).
PowerRequired(3G x N phones) >> PowerRequired(One 802.11 AP).
SpeedAndReliability(One 802.11 AP) > SpeedAndReliability(3G x N phones (N < 20 probably)).
And most importantly:
Cost(N
Old Hat for Cradlepoint (Score:1)
If only I could (Score:2)
You can! (Score:2)
Bandwidth isn't the problem. (Score:3, Informative)
It's not the bandwidth. It's the latency.
Ping on a cell connection runs around 200 ms, in my experience. *That's* the part that makes tethering suck -- with pages requiring dozens of images and javascript files these days, waiting for a 200ms round trip for each request adds up FAST.
Re: (Score:2)
Only if you have a browser that isn't capable of pipelining requests. Which most can these days.
Re: (Score:2)
Verizon Rev-A EVDO does a bit better (most of the time) here:
$ ping 4.2.2.6
PING 4.2.2.6 (4.2.2.6) 56(84) bytes of data.
64 bytes from 4.2.2.6: icmp_seq=1 ttl=46 time=115 ms
64 bytes from 4.2.2.6: icmp_seq=2 ttl=46 time=106 ms
64 bytes from 4.2.2.6: icmp_seq=3 ttl=46 time=92.3 ms
64 bytes from 4.2.2.6: icmp_seq=4 ttl=46 time=111 ms
64 bytes from 4.2.2.6: icmp_seq=5 ttl=46 time=90.6 ms
64 bytes from 4.2.2.6: icmp_seq=6 ttl=46 time=94.7 ms
64 bytes from 4
Instant WiFi access? (Score:2)
I can see it now, people will buy phones to make instant WiFi access to download illegal stuff, and then kill it as quick as it was created, leaving little tracebility to who downloaded what. And this could then put people at risk for being responsible for illegal downloads that they may not have actually had a hand in.
Just a what if thought...
Don't get your hopes up. (Score:2)
Microsoft Research is an awesome entity that produces a ton of cool things. Have you ever poked around on their website? It's got a ton of cool projects like this one. I wouldn't get your hopes of ever seeing this turn into a real product - just because somebody in MS Research is working on it doesn't mean that Microsoft has any plans to use it for anything. I'm convinced that the primary purpose of MS Research is to employ people so that they don't go work for Microsoft's competitors.