Introduction An LDP-based MAC address withdrawal mechanism is specified in [RFC4762] to remove dynamically learned MAC addresses when the source of those addresses can no longer forward traffic. This is accomplished by sending an LDP Address Withdraw Message with a MAC List TLV containing the MAC addresses to be removed from all other Provider Edge nodes over the LDP sessions. [RFC7361] describes an optimized MAC withdrawal mechanism that can be used to remove only the set of MAC addresses that need to be relearned in H-VPLS networks. [RFC7361] also describes optimized MAC withdrawal operations in PBB-VPLS networks. A PW can be signaled via the LDP or can be statically provisioned. In the case of a static PW@ an LDP-based MAC withdrawal mechanism cannot be used. This is analogous to the problem and solution described in [RFC6478] where a PW OAM (Operations@ Administration@ and Maintenance) message has been introduced to carry the PW status TLV using the in-band PW Associated Channel. In this document@ we use a PW OAM message to withdraw MAC address(es) learned via a static PW. Thus@ MAC withdraw signaling for static PW reuses the following concepts: - in-band signaling mechanisms used by static PW status signaling and - MAC withdrawal mechanisms described by [RFC4762] and [RFC7361]. MAC withdraw signaling is a best effort scheme. It is an attempt to optimize network convergence by reducing blackholes caused by PW failover for protected PWs. The protocol defined in this document addresses possible loss of the MAC withdraw signal due to network congestion@ but does not guarantee delivery@ as is the case for the LDP-based MAC withdraw signaling. In the event that MAC withdraw signaling does not reach the intended target@ the fallback to MAC re-learning due to bi-directional traffic or as a last resort aging out of MAC addresses in the absence of frames from the sources@ will resume the traffic via new PW path. Such fallbacks would cause temporary blackouts but does not render a network permanently unusable.