mysql select query slow down exponentially as table grows

I have the following query running on mysql 5.6 on my CENTOS server

select tb1.field2, tb1.field1, tb2.field1 
from dns.table1 tb1, dns.table2 tb2 
where tb1.field2 = tb2.field2 
  and tb1.field1 != tb2.field1

All relevant fields are indexed. on tables of 10 million records, it finishes within 400 econds, however, on 100 million records, it finishes within 10 hours!

Why? I suspect file swapping within mysql, what can be optimized on mysql? would increase buffer pool help?