Why might you want to synchronize IGP and BGP convergence by advertising an infinite metric in OSPF or setting the overload bit in IS-IS?
Click on the arrows to vote for the correct answer
A. B. C. D.A.
In a network environment where both Interior Gateway Protocols (IGPs) and Border Gateway Protocol (BGP) are used, it is important to ensure that their convergence times are synchronized to avoid network instability and traffic loss.
When an IGP experiences a topology change, it can take some time to converge and update its routing tables. During this time, BGP may have already learned about the new path and could start advertising it to its peers, potentially causing routing loops or blackholes until the IGP updates its routing tables to reflect the change.
To avoid this scenario, two techniques are commonly used to synchronize the convergence of IGP and BGP:
Advertising an infinite metric in OSPF or setting the overload bit in IS-IS: This technique involves increasing the metric of OSPF or IS-IS routes to infinity for a specified period of time after a topology change. During this time, the router will continue to advertise the old routes to its peers but with an infinite metric, effectively blocking traffic from being routed through them. This gives the IGP enough time to converge and update its routing tables before the router starts advertising the new routes with a finite metric.
Using route dampening: This technique involves suppressing BGP updates for a certain period of time after a topology change to prevent BGP from converging faster than IGP. When a route changes, BGP will not immediately withdraw the old route and advertise the new one. Instead, it will wait for a certain amount of time before updating its peers, and the duration of the wait will increase exponentially with subsequent changes to the same route. This can prevent BGP from flapping or oscillating between multiple available routes.
Option A is incorrect because advertising an infinite metric or setting the overload bit does not prevent traffic loss when the path from a router to a BGP next hop traverses another router that has not yet learned the path via BGP. In fact, it can potentially cause traffic loss by blocking the old routes temporarily.
Option B is partially correct, as preventing BGP from converging faster than IGP can indeed prevent temporary routing loops in the network. However, the statement that IGP convergence is slower than BGP is not always true and can depend on various factors such as network size, topology, and configuration.
Option C is incorrect because the synchronization techniques mentioned above already prevent routes learned from both IGP and BGP from forming a routing loop.
Option D is partially correct, as using route dampening can indeed prevent churning between multiple available routes reachable through IGP and BGP. However, the use of an infinite metric or the overload bit does not directly address this issue.