BeWifi lets you steal your neighbor’s bandwidth when they’re not using it
arstechnica.com
arstechnica.com
Maybe this makes sense if you're lucky enough to have FiOS (which gives you a dedicated line). For me (a Time Warner customer), this is already the case. Except, it's not "borrowing" so much as it is "stealing".
Technically, I should have 2.5MB/s down (20Mb/s is what I pay for), but in reality, I'm often lucky to get 1MB/s, because my neighbors are using the Internet and watching Netflix, etc. And I have never observed more than 2.5MB/s, so it doesn't average out.
Theoretically, this should already be illegal, but it's still very common in practice.
For this inconsistent connection (and no other services!), I am forced to pay over $50/month.
Edit: Updated - I pay for 20Mb/s, not 15Mb/s.
They run some of the above tests once every hour. The good thing with this initiative is it attempts to build an IETF standard [2] that can be incorporated by all ISPs to make sure that their deployed network is performant.
[1] http://www.samknows.com/broadband/ [2] https://datatracker.ietf.org/wg/lmap/
FiOS uses dynamic bandwidth allocation to enable oversubscription on a segment; therefore, FiOS spends can be impacted by the usage of your neighbors just as cable can be.
So (@dangrossman) the legal issue is averted: both costumers are in the same isp which, to them, works like a black box. So the article's use of the word "steal" is quite nonsensical.
Also remarkable is the advanced anticipation of demand. For example, Netflix is not even available in Spain and bandwidth demand is currently low compared to say the US. My guess is that it is not to facilitate video over the internet etc. but that it is a cheap way to undercut competition from Vodfone and Orange (other Spanish providers). Bandwidth sharing might mean more consistent high bandwidth for those who need it.
Actually the potential issue is inverted: you may not be able not opt out of the system.
The other potential issue is Telefonica itself. Historically it has been as appalling company to customers. It is possible this is changing but there certainly used to be a long track record of awfulness. Indeed it is my utter shock that Telefonica of all companies on earth would manage to do something of this sort that is compelling me to ramble on like this.
If Telefonica is changing, kudos!
Unfortunately, I've never noticed anyone using it. Anecdotally, however, take this as one data point that I've never had a problem with bandwidth hogs.
It sure comes in handy when guests come over---there's no password to explain.
I wish more people would do what you did.
I mean, Skype's going to be sending my conversational partner's audio to a single IP address that represents me, no? How could that be divided among multiple people's bandwidth, technically speaking?
My concern would be my neighbor downloading movie torrents on my connection while I'm asleep. We'd need a really strong legal precedent that the person paying for the connection isn't responsible for what others do on it before this kind of thing could catch on in most countries.
But, the devil's in the details.
Again though, without knowing more about the protocol/mesh they are using, it would be impossible to know.
So the mesh stuff has to happen behind NAT, on local IPs.
In almost every product like this I've seen, it's not load-sharing a single session (like a Skype call) across multiple Internet connections, but instead sharing/routing particular sessions to particular networks.
Skype for instance is relatively low-bandwidth, but latency sensitive. Downloads are high bandwidth, and less concerned with latency (within reason). The router can make decisions, most likely based on the protocols and/or ports of the session, and route each session to the most appropriate outbound Internet connection.
You can aggregate multiple pipes into one larger "bonded" connection, but this would generally require equipment on both ends of that connection. Technically, feasible, but most likely not in this case.
They could also use the mesh network to route some packets, or segments (eg: bunch of packets in a row) to the best outbound pipe, which would be selected based some continuous ping tests (in essence). You could probably make that happen with transparency to the end user, even in Skype.
In the end, from the very limited info, this sounds like the next rev of basic multi-homed "load sharing" routers that have been around forever, just with a new twist.
Unfortunately it's difficult to achieve this sort of innovation on a large scale, because we don't have enough IPv4 addresses, and everyone's dragging their feet on the IPv6 upgrade.
I also wonder if HBO and other channels might get overzealous and shut people down for the p2p nature of it.
But all that said, this is the obvious future of the web until we have a true, free, ad hoc/distributed network and on some level I can't believe it took this long to get the ball rolling..
http://tools.ietf.org/html/draft-singh-avtcore-mprtp (for skype and multimedia)
https://tools.ietf.org/wg/mptcp/ (for everything else, even Siri)
Regardless it looks like these guys are just taking the Commotion firmware and re-branding it as a service. It also doesn't solve the local node problem. Most people don't understand how cable (HFC), fiber and copper (DSL) are built out to the neighborhood or 'node'. It's all oversubscribed and shared. So this type of solution really does nothing for the local congestion in a node. To me, this is a pure marketing gimmick.
OTOH, if the last mile is oversubscribed, there is very little one can do.
Anyone can connect to it for free if you're a BT customer elsewhere or you can pay. It doesn't effect my bandwidth as most plans are unlimited anyway.
http://wifinetnews.com/archives/2004/06/radiuz_combines_wpa-...
Perhaps this solves the problem of not having to deliver as much aggregate bandwidth as isolated pipes on a per carrier basis, but cross carriers may or may not go for it, like "roaming."
http://arstechnica.com/information-technology/2013/06/comcas...