Across a large WAN network, there will be new video traffic being distributed from a single source at any given time however, the video source might originate from different parts of the multicast…

Any Source Multicast relies on (*,G) shared trees and does not inherently guarantee optimal path selection to a specific source since the shared tree is anchored to an RP rather than the source.

PIM Sparse Mode routes initial traffic through a Rendezvous Point that may not be on the optimal path, and SPT switchover requires additional convergence time whenever the source location changes.

Bidirectional PIM uses a shared tree rooted at the RP in both directions and never switches to a source-specific Shortest Path Tree, making it structurally incapable of providing optimal per-source path selection.

SSM creates (S,G) multicast state that constructs a Shortest Path Tree anchored at the actual source address, ensuring receivers always receive traffic via the most efficient path to that source. Because the tree is always source-specific and shortest-path-rooted, changing source locations simply means building a new optimal SPT for the new (S,G) pair without ever routing through a suboptimal Rendezvous Point.