one of the most expensive operations is loading from a snapshot - specifically, generating the account index - ideally if you do it once, you should be able to re-use the index created and load much faster

this will be split up into PRS:

• improved naming/organization (feat(accountsdb): improved struct/var naming #320)
• changes to the account index to support fast loading
  • improved allocation approach (feat(accounts-db): support fast loading #324)
  • index instead of ptr (feat(accounts-db): change reference pointers to indexes #358)
• code which saves and loads the index data

note: old branch with more context can be found in #303