Why use Http instead of Curator to find segments and loadQueue in Druid 0.12.0

I think using zk to find segments looks like good way. What advantages of using http instead of zookeeper ?