The example database/server provided by Apple for Live Caller ID contains a hardcoded database with a tiny number of pre-defined numbers.
However, its not expected to be representational of an live real world usage server.
But the question is how can that be accomplished if its a requirement that the data be KPIR encrypted?
In real world scenarios, the factors that effect whether a number should be blocked or not are continually changing and evolving on a minute-by-minute basis and new information becomes available or existing information changes. If the database supports tens of millions or hundreds of millions of constantly changing phone numbers, in order to meet the requirements of the Live Caller ID being KPIR encrypted, that would imply the database has to re-encrypt its database of millions endlessly for all time. That seems unfeasable and impractical to implement.
Therefore how do the Apple designers of this feature envisage/suggest a real-world server supporting millions of changing data should meet the requirement to be KPIR encrypted?