These two techniques both influence the likelihood of a requested object being memory resident.
The first component directs each object request to a nearby ad-hoc node that has the requested object, while avoids overloading any of the nodes.
A client must also pass a key, for example a specific type, to identify the requested object.
If, on the other hand, the requested object is not cached, the system gets it from the origin server before returning it to the requesting user.
Each log entry contains information about a single user request, including the identification of the user that made the request (abstracted as a unique numeric identifier for privacy reasons), the id of the requested object (also a unique identifier), the timestamp of the request, the number of bytes of the requested object and the HTTP code of the response sent to the user.
In addition, a cache miss triggers the eviction phase of the algorithm if the requested object does not fit in the available free space.
On a cache hit, the frequency and recency information of the requested object is updated and its cached version is returned to the user.
Implicitly, this metric is related with the probability that a user who requested object x i will visit another object x j to which there may be a navigation path rooted in x i.
In the handle-driven mode, it returns a reference for the requested object to the client that uses this reference to re-send the request for the desired operation to the server, whereas in the forwarding mode it forwards the entire client request to the server.
Then, all objects of the requested object type are requested within the search window, the given descriptor is verified with the descriptors of the collected objects, and finally that object in the search window is selected which matches the descriptor [7].
