When the router or system auto-creates a bypass LSP at runtime, these bypass LSPs are called dynamic bypass LSPs. Dynamic bypass LSPs are created at a FRR LSP PLR when there is a requirement to provide FRR facility backup protection. At every PLR, the facility protected FRR LSP requiring backup can trigger the creation of a dynamic bypass LSP.
A dynamic bypass provides an automatic creation of a bypass LSP protecting any MPLS interface, based on the user configuration. Unlike static bypass LSPs, dynamic bypasses protect only one interface at a time.
The backup path of FRR LSPs give static bypasses preference over dynamic bypasses. When there are existing static bypasses which can satisfy the backup constraints, then static bypasses are used; otherwise, dynamic bypass creation is initiated when the system configuration allows.
When establishing a facility protected LSP with link or node protection, each LSR on the primary path verifies if there are any existing bypass LSPs that satisfy backup path constraints. Backup constraints include MP, backup bandwidth, hop-limit, priority, exclude links, and excluded nodes. The protected LSP backup path uses this bypass to reach its merge point. When there is no bypass available, the PLR node computes and establishes a new bypass LSP, addressing the backup path constraints.
When creating a bypass LSP for an interface, any number of facility protected LSPs may share the same bypass LSP, providing it satisfies the FRR backup constraints. Adaptive dynamic bypass LSPs can also be re-optimized periodically using the make-before-break procedure. A make-before-break switch from the current instance of a bypass to a new Instance releases all the backup paths it is attached to. All backup LSPs again perform a re-query of the backup path bypass search.
Dynamic bypass global and interface level configuration parameters, such as bandwidth, hop-limit, and priority are used while creating a new dynamic bypass LSP. Any modifications on these global and interface level configuration parameters are taken into consideration during the next cycle of re-optimization or during manually initiated bypass LSP re-optimization.
Backup path optimization does not happen while a backup is already using a bypass LSP. For the backup path to move to an optimal bypass LSP, the backup path must relinquish the current bypass LSP and go through the backup query process again. This query process searches for an optimal bypass LSP for the backup.
The bandwidth of the newly triggered bypass LSP is zero by default unless it has an explicit configuration at interface level or the backup has the backup bandwidth request. When a new facility protected LSP requests a bandwidth which cannot be accommodated within an existing dynamic bypass LSP, there is no automatic make-before-break for the existing dynamic bypass LSP. Instead, a new dynamic bypass is created, depending on the configurations and system limits.
Similar to static bypasses, backup paths from multiple facility protected LSPs can use dynamic bypasses providing they share the same protected-interface, merge point, and satisfy FRR backup constraints.