Of course this is a quick and fast solution for bhuiraj, it's not optimal, but do not require special software. I've tested 1.5Gb of data for over 227 millions of words, and the DB is quite big and the search are not so fast. But, of course, if you need speed you can use DB like mysql or oracle using a fine tuned configuration (memory index, cache query, partition table, etc.).
In this case, however, is possible create an optimal solution (without the generic DB overhead), but you need to create a specific software to handle a very very big dictionary.