BTW, I agree with @MHM Cisco World , your question probably should have been its own posting.
In any case, VRFs are sort of the equivalent of VLANs for L3.
Quick review, given 3 routers, linearly connected, with loopbacks 1.1.1.1, 2.2.2.2 and 3.3.3.3, and these loopbacks were known to OSPF, i.e. R1.1.1.1<>R2.2.2.2<>R3.3.3.3, if there was just one area number (doesn't need to be area zero) for all 3 routers, each router knows of all the routes within the OSPF AS.
If we go multi area, having area 1 shared between R1.1.1.1 and R2.2.2.2 and area 3 between R2.2.2.2 and R3.3.3.3, R2.2.2.2 would still know all routes, but R1.1.1.1 and R3.3.3.3 wouldn't know the other area's routes, unless we add an area zero on R2.2.2.2, and if we do, sharing of routes between areas would depends on the kind of area and/or if route summarization is being done on ABRs. Even in the latter case, usually we can still route between all the areas even without knowledge of external area routes.
Next, on R2.2.2.2 we could run two OSPF processes. Those processes would still allow R2.2.2.2 to know all routes, but again, R1.1.1.1 and R3.3.3.3 would not, unless we redistributes between those OSPF processes on R2.2.2.2.
In the forgoing, we have the "problem" R2.2.2.2 might have more routing info than we desire. I.e. We truly want to logically separate R1.1.1.1<>R2.2.2.2 from R2.2.2.2<>R3.3.3.3. This is where VRFs come in. We can define two VRFs on R2.2.2.2, and whatever routing information is known in one VRF, by default, is not available to the other VRF.
Two issues that arise with VRF, how do we share multiple VRFs across/between L3 devices (the equivalent of sharing a VLAN across/between switches) and is there a way to selectively "redistribute" routes between VRF.
On a single device, we can identify interfaces or static routes to a VRF, much like assigning a L2 port to a VLAN.
On L3 devices, we can identify SVI to a VRF, and implicitly, we can use a trunk (with VLAN tagging), to share a VRF across/between them. This is the VRF-Lite approach.
Full blown VRF, uses VLAN like tagging (RDs) of BGP routes to identify what VRF they belong to. (Unsure any other routing protocols support that too.)
Full blown VRF can use RTs to selectively "filter" BGP routes, on a particular device, so there is a way "leak" routes between VRFs. (One reason to do this, might be to allow separate VRFs to share an address space. [Sort of the equivalent of private IPs seeing public IPs, but not the converse.])
Again, I don't know if any other routing protocols support RD/RTs as can BGP.
If any of the above makes any sense, to your questions, VRF-Lite, basically allows you to run totally independent OSPF ASs on the same router, and those independent ASs can be shared across multiple devices, again, conceptionally, much as having multiple VLANs on a single switch and/or sharing multiple VLANS between switches. VLANs, of course, isolate L2, while VRFs isolate L3.
In the Enterprise, you often don't see much need for VRFs, unless you're very, very security conscience. Consider you have "sales" and "accounting" and you want to preclude them from directly being able to exchange data, except perhaps through a FW. With VRF, you can route these two ASs across your network, but each has no way to reach the other, except perhaps, through a FW.
Many VLAN examples might show "sales" and "accounting" VLANs, VLANs might have L2 size and topology considerations which can be better addressed by L3 using VRFs. Remember, even on any L3 devices with different VRFs, no need to maintain ACLs to block traffic between the VRFs.
