21 points | by mmaia 3 days ago
5 comments
MaxMind also significantly reduced their MMDB file sizes recently, on average by about 30% as well https://dev.maxmind.com/geoip/release-notes/2026-10-21-signi...
> The search tree is a binary trie over the bits of an IP address.
The interesting thing here is that the format is still the same.
I read the original spec and it did call out that
> In practice, data deduplication is handled by pointers, making this type unnecessary.
If any entry can be a pointer, you can shrink this recursively for each field type like a .strings table design.
These are widely deployed [citation needed] so nice chance to save some ram!
I wonder - if the number of nodes goes down enough then they could reduce the number of bits for the record_size and save another 12-25%
this article would probably be awesome if it wasn't AI written. It's terrible to read
AI is also how the made it 33% smaller.
MaxMind also significantly reduced their MMDB file sizes recently, on average by about 30% as well https://dev.maxmind.com/geoip/release-notes/2026-10-21-signi...
> The search tree is a binary trie over the bits of an IP address.
The interesting thing here is that the format is still the same.
I read the original spec and it did call out that
> In practice, data deduplication is handled by pointers, making this type unnecessary.
If any entry can be a pointer, you can shrink this recursively for each field type like a .strings table design.
These are widely deployed [citation needed] so nice chance to save some ram!
I wonder - if the number of nodes goes down enough then they could reduce the number of bits for the record_size and save another 12-25%
this article would probably be awesome if it wasn't AI written. It's terrible to read
AI is also how the made it 33% smaller.