I prefer to use storable and place big data structures into a storable file. It's a bit like do/require FILE but easier and faster but i suppose the main advantage is it's ability to compact data by freezing it for later use or shared data in inter-process communication.

There's also the Perl Data Language which allows one to efficiently store data and in the process make these operations faster. It is not readily intuitive and has a high learning curve but definitely worth considering.

update: removed question about pickle files.