I'm wondering does Cypher (runtime) apply the traversal framework (The traversal framework - Java Reference)?
Or they are actually two different ways of accessing the graph? (which means their implementations are totally different)
And if they are actually different, I'm wondering normally which one will be better for simple queries?
The traversal API is very different from Cypher. The primary difference between the two is that Cypher is declarative while the Traversal API is imperative.
For simple queries, Cypher is better. The traversal API is only faster/easier in certain specific cases, none of which are usually described as "simple."