What is SAVI
SAVI (Source Address Validation Improvement) is a security mechanism used on access devices to validate the authenticity of IPv6 Neighbor Discovery (ND) packets. It is built on ND Snooping, DHCP Snooping, and static binding entries, and it helps prevent unauthorized packets from entering the internal network.
In an IPv6 access network, hosts rely on ND messages to find neighbors and reach the gateway. Without validation, a user who manually configures a private or forged IPv6 address can send ND packets as if it were a legitimate host, and may be able to impersonate other devices or gain access it should not have. SAVI addresses this by checking every ND packet against the address bindings the switch already trusts, so only hosts with a verified address and MAC pairing can communicate through the access port.
How SAVI Works
SAVI validates users by comparing the source addresses in ND packets with three types of trusted records held on the device:
Static binding entries: Bindings configured manually by the administrator.
ND Snooping entries: Bindings learned by snooping ND packets.
DHCP Snooping security entries: Bindings learned from DHCP address assignment.
The validation logic is straightforward:
A host sends an ND packet into a VLAN where SAVI is enabled.
The device checks whether the packet's source IPv6 address and source MAC address match an entry in any of the three records above.
If a match is found, the packet is considered legitimate and is forwarded. If no match is found, the packet is treated as unauthorized and discarded.
Two exceptions are worth noting. Router Advertisement (RA) packets are not subject to SAVI checks and are controlled by the RA Guard function instead. Router Solicitation (RS) packets with a link-local source address are allowed to pass without further validation.
SAVI is enabled per VLAN. Ports can also be configured as SAVI trusted ports: when SAVI is enabled for a VLAN, ND packets received on a trusted port with that VLAN ID skip SAVI validation and are allowed to pass. This is typically used for uplinks and other links toward trusted devices.
Why SAVI is Beneficial
SAVI adds source-level protection to the access layer without changes on end hosts:
Blocks unauthorized users: Hosts using private or forged IPv6 addresses cannot get their ND packets through, so they cannot access the network.
Strengthens the IPv6 access edge: Validation happens at the first switch the packet reaches, before it can affect other devices on the network.
Reuses existing binding information: Checks rely on ND Snooping, DHCP Snooping, and static binding entries, working together with other access-security features rather than adding a separate database.
Simple to operate: Only one required task, enabling SAVI on the VLAN, with trusted ports as an optional refinement.
Clear division of roles: SAVI handles ND packet validation, while RA Guard controls Router Advertisement packets, so each threat is handled by the right function.