Wait, you're saying Oracle (and Sybase) are not natively relational? Do tell more...
Assuming that you are not just a bad troll looking to feel important . .
Why do Oracle and SyBase engines use indexes?
Because indexes provide a traversable tree structure that "terminate" with direct access addresses for the individual records within the system.
Maybe you are unaware that the existence of indexes is not part of the RDBMS definition?
Why does the primary index definition and implementation affect the performance of an Oracle or SyBase instance? Because the records are physically sorted into the order of the primary index in order to make traversing the contents of the database in that order a trivial issue.
Maybe you are unaware that RDBMS theory mandates that it is impossible to predict the order of data within a query response unless the query mandates an order?
Why is so much of the overhead associated with running both Oracle and SyBase dedicated to maintaining indexes? Because neither system has the ability to provide a reasonable query return time when you ask for a response set in a non-indexed order.
Maybe you are unaware that RDBMS theory mandates that the delay in a query response should relate to the complexities of the joins used in the query, not the requested order of data?
Or were you looking for something more?