Sham Links vs. Virtual Links?(VL)

Well, several weeks ago, Ceta joined an orkut community in which talking about OSPF networking. One of the topics is discussing the different between sham links and virtual links in OSPF network implement or designing. ha, in fact, they're totally different solutions for different networking design, so it's not comparable between SL and VL. I should reply the post except that there's handicap in posting figures on orkut. So, let me explain the detail below, and wish "Pankaj" will see my blog here.
-------------------------------------------------------------
Can anybody give me some idea about the difference between OSPF sham link & virtual. I mean in which scenario sham link is used & in which virtual link is used. What LSA type is flooded over a sham link & what LSA type is flooded over virtual link.
-------------------------------------------------------------
Virtual Links comes the first part: All areas in an OSPF AS must be physically connected to the backbone area (area 0) or area 0.0.0.0 if you configure your network on Juniper Routers. Unfortunately in some cases where this physical connection is not possible, you can use a virtual link to connect to the backbone through a non-backbone area. At the same time, you can also use virtual links to connect two parts of a partitioned backbone through a non-backbone area. So, there’re two functions of VL in all. I emphasize here because so many engineer ignore the second one in their work, and this may cause a sub optimized solution which created two different OSPF processes for each backbone area, then make an inter-process redistribution in order to received complete route from both sides. Further more, the area through which you configure the virtual link, known as a transit area, must have full routing information. The transit area cannot be a stub area. In this example, the virtual link connects area 7 to the backbone through area 5:

A virtual link is created between the routers with router ID 1.1.1.1 and router ID 2.2.2.2:
-------------------------------------------------------------

R1(config-router)#area 5 virtual-link 2.2.2.2
R2(config-router)#area 5 virtual-link 1.1.1.1
-------------------------------------------------------------
After the VL has been established, routes from area 7 should be transit to the backbone area by the VL, vice versa. Now, I would like to do a further explanation about VL here, because "Pankaj" mentioned about the LSA type in his post. The question is “what is a Virtual Link in fact?” or “can we solve the problem without a virtual link technology? Let’s disconnect the virtual link first:
-------------------------------------------------------------
R1(config-router)#no area 5 virtual-link 2.2.2.2
R2(config-router)#no area 5 virtual-link 1.1.1.1
-------------------------------------------------------------
As soon as you issue this commands, the routes from area 7 now disappear on the backbone routers. Then, we create a tunnel between R1 and R2, and we put the tunnel in the OSPF backbone area, let’s see what will happen:
-------------------------------------------------------------
R1(config)#interface tunnel 0
R1(config-int)#ip address 10.1.1.1 255.255.255.0
//allocate an IP address for this virtual interface
R1(config-int)#tunnel source loopback 0
R1(config-int)#tunnel destination 2.2.2.2
//make sure there’s useable route to the destination
R1(config-int)#router ospf 64
R1(config-router)#network 10.1.1.0 0.0.0.255 area 0
-------------------------------------------------------------
R2(config)#interface tunnel 0
R2(config-int)#ip address 10.1.1.2 255.255.255.0
//allocate an IP address for this virtual interface
R2(config-int)#tunnel source loopback 0
R2(config-int)#tunnel destination 1.1.1.1
//make sure there’s useable route to the destination
R2(config-int)#router ospf 64
R2(config-router)#network 10.1.1.0 0.0.0.255 area 0
-------------------------------------------------------------
Now, a tunnel is created between R1 and R2, then we find the flow instances:
1. R1 and R2 establish a new neighbor ship with each others, by the interface “tunnel 0” which is the virtual interface we created above.
2. R1 changed it's role from an internal router to both of an ABR and a Backbone Area Router. Note, R1 was not an ASR before, though it connects both area 5 and area 7. The router which connects backbone area and other area(s) is defined as an ABR, a router will not be defined as an ABR unless it connects the backbone area, no matter how many non-backbone area it has connected.
3. The most important, all route generated from area 7 now has been advertised to the backbone area now, vice versa.

I think we could make a summary here, a virtual link actually is a tunnel which fall in the backbone area, and expand the backbone area to a router which is not connected directly to the backbone area, then make the isolate router an ABR. Only ABR would generate LSA3 to the the backbone area. So, OSPF floods LSA1 in the tunnel or we call it virtual link customarily to build the neighbor ship and make itself an ABR, then flood LSA3 to the backbone. I will talk about sham-links next time.