> > Basically, we get bulk data from the five RIRs, and try to parse > > it into a structured database. For the ones like ARIN and LACNIC, that > > are coming out of a similar relational database, it's not too difficult > > work. For others, it's largely manual. > > Its not 100% clear for LACNIC & ARIN either. ARIN & LACNIC do allow for > common organization that is a db parent for both ASN and an ip block but > there are actually number of organizations with multiple Org IDs and with > ASN being under different org (also same same sometimes happens for IPv6 > allocations) and cases of organizations buying each and ending up with > multiple ones (but only using one asn) are common. Yes, we've got a table of organization-to-OrgID, particularly since one organization will have different OrgIDs not only within one RIR, but in different RIRs, even. My only point was that it's a lot easier to get data back into a database if it came out of a database in the first place. Then all you have to do is understand the relationship model between the objects, and understand what limitations it has and how it maps into one's own superset of that. -Bill