The maximumpaths Command and BGP Decision Process Tiebreakers

The goal of the BGP decision tree is to find the one best BGP route to each NLRI, from that router's perspective. That router then considers only its best routes for advertising to other routers, restricting those routes based on AS_PATH loop prevention and routing policy configuration. That router also attempts to add that best route, and that best route only, to its IP routing table. In fact, as long as another routing source has not found a route to the same prefix, with a better administrative distance, the best BGP route is placed into that router's routing table.

If BGP has not chosen a best route for a particular NLRI after Steps 0 through 8, then multiple routes tie for being the best route. At this point, BGP needs to make two important decisions:

■ Which route is best—BGP uses two tiebreakers, discussed next, to determine which route is best.

■ Whether to add multiple BGP routes for that NLRI to the IP routing table—BGP

considers the setting of the maximum-paths command to make this decision, as described after the discussion of Steps 9 and 10.

Even if BGP adds to the IP routing table multiple BGP routes to the same prefix, it still picks only one as the best route in the BGP table.

Step 9: Lowest BGP Router ID of Advertising Router (with One Exception)

The first tiebreaker is to pick the route with the lowest RID. The logic is actually two steps, as follows:

1. Examine the eBGP routes only, picking the route advertised by the router with the lowest RID.

2. If only iBGP routes exist, pick the route advertised by the router with the lowest RID.

These straight-forward rules are followed in some cases, but not in some others. The exception to this rule occurs when BGP already has a best route to the NLRI, but it has learned new BGP information from other routers, including a new BGP route to reach a previously known prefix. The router then applies its BGP decision process again to decide whether to change its opinion of which route is best for that NLRI. If the decision process does not determine a best route by this step, this step uses the following default logic:

If the existing best route is an eBGP route, do not replace the existing best route, even if the new route has a smaller RID.

The reasoning is that replacing the route could result in route flaps, so keeping the same route is fine. This behavior can be changed so that the lowest RID is always used, by configuring the bgp bestpath compare-routerid BGP subcommand. Note that this exception only applies to eBGP routes; if the currently best route is an iBGP route, the decision is simply based on the lowest advertising router's RID.

Step 10: Lowest Neighbor ID

If Step 9 did not break the tie, then the router has at least two neighbor commands that point to the same router, and that router happens to have the lowest RID of all current neighbors advertising the NLRI in question. Typically, if redundancy exists between two routers, the configuration uses loopback interfaces, a single neighbor command, and the neighbor ebgp-multihop command if the neighbor is an eBGP neighbor. However, using a pair (or more) of neighbor commands pointing to a single neighboring router is a valid configuration option; this final tiebreaker provides a way to break ties for this case.

At this point, the router looks at the IP addresses on the neighbor commands corresponding to all the neighbors from which the route was received, and it picks the lowest neighbor IP address. Note that, as usual, it considers all routes again at this step, so it may not pick the neighboring router with the lowest RID at this point.

The BGP maximum-paths Command

BGP defaults the maximum-paths command to a setting of 1—in other words, only the BGP best route in the BGP table could possibly be added to the IP routing table. However, BGP will consider adding multiple entries to the IP routing table, for the same NLRI, under certain conditions— conditions that differ based on whether the best route is an eBGP route or an iBGP route.

First, consider eBGP routes. The following rules determine if and when a router will add multiple eBGP routes to the IP routing table for a single NLRI:

KEY 1. BGP must have had to use a tiebreaker (Step 9 or 10) to determine the best route.

POINT

2. The maximum-paths number command must be configured to something larger than the default of 1.

3. Only eBGP routes whose adjacent ASNs are the same ASN as the best route are considered as candidates.

4. If more candidates exist than that called for with the maximum-paths command, the tiebreakers of Steps 9 and 10 determine the ones to use.

Although the list is detailed, the general idea is that the router can trust multiple routes, but only if the packets end up in the same adjacent AS. Also, BGP must restrict itself to not use multipath if the best route was found via Steps 0 through 8 of the decision process, because forwarding based on another route could cause loops.

Next, consider iBGP routes. The rules for iBGP have some similarities with eBGP, and a few differences, as follows:

KEY 1. Same rule as eBGP rule 1. POINT

2. The maximum-paths ibgp number command defines the number of possible IP routes, instead of the maximum-paths number command used for eBGP.

3. Only iBGP routes with differing NEXT_HOP settings are considered as candidates.

4. Same rule as eBGP rule 4.

The rationale is similar to eBGP with regard to most of the logic. Additionally, it does not help to add multiple IP routes if the NEXT_HOP settings are equal, so BGP performs that additional check.

Finally, the maximum-paths eibgp number command seemingly would apply to both iBGP and eBGP routes. However, this command only applies when MPLS is in use. Table 13-14 summarizes the key commands related to BGP multipath.

Table 13-14 BGP maximum-paths Command Options

KEY POINT

Table 13-14 BGP maximum-paths Command Options

Command	Conditions for Use
maximum-paths number	eBGP routes only
maximum-paths ibgp number	iBGP routes only
maximum-paths eibgp number	Both types, but MPLS only

Continue reading here: BGP Communities

Was this article helpful?

+4 0