MicroMap files are organized in both horizontal layers (regions) as well as vertical layers (features).
Large geospatial data like vector maps require a server-side decomposition into smaller data chunks. By default we divide the world map in many non-overlapping regions, preferably by continents, countries and even further down the hierarchy. For a hybrid solution we successively divide those regions in even smaller tiles until we reach a well manageable file size, with a target of let's say 100 kbytes. MicroMap maps are created automatically from UMDB, our unified mapping database technology which supports literally any vector-based map. Since MicroMap comes with an API developers can easily implement spatial or map based applications without dealing with vector map data directly.
The vertical layers are by content type and/or data source and each layer can have different latency periods and update frequencies, this can mean (depending on the specific layer) that we indicate that the cache of that layer should be removed after a certain period of time or that its normal lifetime is up to a certain period of time so that the application can make a determination if the data should be trusted. The lowest priority in any sync would be data blocks that have changed but are inside their lifetime.