BGP information for customers

AS5539 currently does not carry unfiltered ``full BGP'' tables, due to the large growth of the global table.

The global table IPv4 is filtered as follows:

• Prefixes coming from SpaceNet PA blocks or IXPs that we are connected to are blocked.
• From /8 networks assigned to providers in the RIPE NCC service region, all prefixes up to a maximum prefix length of /24 are accepted. /25 to /32 are dropped.
• From other RIR service regions, in those /8 blocks that have a shorter ``minimum allocation size'' (e.g.: the RIR will only ever assign /22s or shorter prefixes), /24 and /23 prefixes are dropped. See here for the current list of filtered /8s, in Cisco prefix-list format.
• To compensate for prefixes that are only announced as more-specifics and have no aggregate route, and would otherwise be not reachable at all, we carry a static default route that points to our upstream providers.

This filtering brings the number of routes we have to carry from about 280.000 down to about 230.000, without significant connectivy reduction.

The global IPv6 table is currently filtered for bogon networks only.

Customer routes, both IPv4 and IPv6, are filtered according to the RIPE NCC database. If you want to announce anything to AS 5539, make sure it is registered there. Due to the low frequency of changes in our typical customer base, this is currently not updated automatically - if you have modified your prefixes, please let us know at bgp@space.net.

Propagation of BGP routing inside SpaceNet is controlled by BGP communities. The specific values used to tag external routed and available for customers to control their routing are documented here.