I have a rather large ontology (about 80 GB). I think the reasoner brings the whole ontology into the memory for reasoning process, doesn't it? is there any way to reason over a 80 GB ontology with 16 GB of RAM?
Before you even start with reasoning the ontology will need to be loaded in full by Protege.
On top of that the reasoner will require further memory to do the actual reasoning. How much memory the reasoner will require is highly dependent on the expressivity of the ontology (i.e., EL++ vs OWL DL), the characteristics of the ontology (i.e., lots of OR branching will require more memory), and reasoner you plan to use. The following paper may be useful in this regard.
It may be worth thinking about the following:
(1) What are the inferences that you hope to derive through reasoning? If only a small number of entailments can be expected it may not be worth the effort.
(2) Is it possible to break up the ontology into smaller modules in such a way that reasoning on the module will give the required entailments?
(3) If your ontology contains a large number of individuals what you may rather need is a triple store on which you can execute rules rather than a reasoner.