On Thu, Mar 29, 2012 at 07:41:17PM +0200, Michael Richardson wrote: > Jeffrey> Let's say my Mother has two internet providers, A and B. > Jeffrey> The reason may be because she needs provider A for home > Jeffrey> work, but it's a somewhat filtered network (she can't get > Jeffrey> to her facebook!) and has B as her personal provider. > > A good example, particularly if "A" is really a virtual wire.
The scenario I have in mind is more of something that appears as a device on the network. On-host VPN clients that don't serve anything off-host seem somewhat out of scope? > But, my question is: on provider A, will she have full access, or only > access to work? It's possible that it may be full access. To me, this really is a matter of "does the work device advertise default?" Some of this is potentially business policy as well. For example, the work link may be fine with the link being used for general Internet access for work purposes and perhaps even goes through a nice firewall to scrub things. However, they may prefer that any personal stuff happens on the home network. A simple example is "no social networking across company resources". > Jeffrey> Instead of source routing, I can also see standard routing > Jeffrey> being done. Presume 192.0.2/24 is where the work-related > Jeffrey> resources are. A static route configured on the gateway > Jeffrey> for A and advertised into the routing protocol would be > Jeffrey> sufficient to sink work traffic to the right router. > > so, this is what we have been calling a walled-garden, but perhaps that > term is too confusing due to WAP-type usage a decade ago, and we need > another term. FWIW, just keep it simple. It's routing. Routing protocols, for home users, may be hard. However, a UI that says "the route to some resource goes through me" is still intelligible for home users as long as the UI is simple enough. -- Jeff _______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
