In the IGP era, we all know, for example, OSPF, link state protocol, in the routing of the time, than the cost of the link, which link consumption will be more priority, if two of the same link consumption, then do load balancing.
In fact, the 2 principles to achieve the OSPF---shortest path first work principle.
What if, in BGP, there are multiple routes to achieve the goal, how should the choice be made? Which path is the preferred one?
Take a look at the priority of these 11 attributes, and deepen the memory, preferably with a skilled back:
1 Highest Weight--cisco private properties
2 Highest local prerence-----recognized optional
3 ROUTE originated by the ROUTER NEXT hop=0.0.0.0 *-----Recognized must be respected well-known mandatory
4 Shortest as PATH *-----recognized well-known mandatory
5 LOWEST originated igp> Egp>incomplete *-----Recognized well-known mandatory
6 LOWEST MED
7 EBGP path over IBGP path
8 prefer the PATH throgh the closest IGP neighbor
9 Rrefer oldest ROUTER for EBGP PATH
Prefer the PATH with the LOWEST neighbor BGP ROUTER ID
One BGP LOWEST ROUTER ID
MINIMUM Claster LIST LENGTH
In BGP, in addition to a large set of principles to be understood, the principle of route selection is the most important. Must be mastered (to understand and troubleshoot BGP, here must be to grasp rather than understand) the route principle.
Each property is validated by the document one at a time. To be able to reach myself in the mind of these concepts.
Let's take a look at the BGP route selection principle:
1. Select the path with the maximum weight (weight) value, the weight is Cisco proprietary attribute.
2. If the weight value is the same, select the route with the highest local priority first.
3. If the local priority is the same, the BGP route on the local route (the next hop to 0.0.0.0) is selected first.
4. If the local priority is the same, and there is no route originating from this router, a route with the shortest as path is preferred.
5, if the same as the length of the path, the first choice with the lowest source code (ibgp< Ebgp<incomplete) route.
6. If the origin code is the same, select the path with the lowest med value.
7, if the Med is the same, in EBGP Routing and Federated EBGP routing, the preferred EBGP route, in the Federation EBGP Routing and IBGP Route, preferred Federation EBGP route.
8, if all the previous attributes are the same, select the nearest path from the IGP neighbor.
9. If the internal path is the same, select the path of the lowest BGP router ID first.
Next-hop attribute Detailed:
I checked the room to Yong IE notes, as well as the Chinese version of the TCP/IP routing Technology (volume II), there are some Cisco training ppt. In this information, three principles are mentioned about next-hop. will be covered in this document. However, I personally think that since it is the principle of the excerpt, it must be compared, and all the information mentioned in the work principle, then my question is, if weight and local_pref are the default situation, the third comparison next-hop how to compare? With the help of the well students, found the English version of the description, only to solve the mystery. I should say that before I check what data are no problem, but are less a key thing, if the comparison next-hop this problem, what is called: if the local priority is the same, priority to select the local route (the next hop to 0.0.0.0) on the BGP route.
Prefer the path that is locally originated through a network or aggregate BGP subcommand, or through redistribution from An IGP. Local paths sourced by Network/redistribute commands are preferred over local aggregates by the sourced Command.
First of all, let's say Next-hop's three working principles:
Principle I:
If the location of the router and the receiving router is declared to be in a different as, then Next-hop is the interface address of the routing.
This is indicated on TCP/IP volume two.
This is the same as IGP. There's nothing to say, if the BGP update message is declared through EBGP, then the Next-hop property will be the IP address of the EBGP.
The ordinary IGP is router by router, and BGP is as by.
So when two ebgp in the notification route, the effect is the same as the IGP effect. Because two routers are ebgp each other, they are ASBR.
So the next hop is just another area's ebgp neighbor address.