CDB did many "bitty" reads/writes when scanning or writing hash tables - change these to bult reads and writes to speed up. CDB also added capabilities to fetch positions and get keys by position to help with iclerk role.