redpanda-data
ntp rework for topic identifiers
not a todo for this commit, but @dotnwat , @rystsov - i wonder if topic_partition and ntp which are really quite heavily used, if we should move them to be value-types say 3 integers and every core has a model::canonical_topic_name, but by default the model::topic_name is really an alias to a int64_t and when you print it, it does a lookup on a core-local table.
Maybe we make a todo if this becomes a problem in perf.
Nothing to do in this comment, but since we manipulate that data type at the heart of the system, probably worth optimizing for.
Originally posted by @senior7515 in https://github.com/vectorizedio/redpanda/pull/1182#discussion_r615335338
Actually kafka is moving to topic identfiers https://cwiki.apache.org/confluence/display/KAFKA/KIP-516%3A+Topic+Identifiers
