Want to create a model for effective memory utilization with faster access

Question

Can someone help me I am looking for a solution to a problem. Problem description is as follows:
We have a data model like:
name
City
Address
Zipcode
1. we have a huge numbesr(millions) of such objects availiable in the memory.How can i make a good design for the better memory utilization.
Means in which structures data should be stored in a memory to make effective memory utilizaion.We already have data structures like hashmap
,hashtable but beyond that can we use them or use other data structures in such a way that memory utilized by these objects is minimal.
2. design should be created keeping in mind that we can apply filters on any of the model attributes.(like if we want to see data of those
objects only where city name is newyork) so filtering done on the data should be fast.

Answer 1

Perhaps you're trying to solve the wrong problem? If the true objective is "to retrieve data as quickly as possible," perhaps you should investigate a database rather than trying to squeeze things into the smallest possible memory footprint? You'd have to have some pretty hefty hardware to keep "millions" of records in memory in addition to applications, server, OS, IP stack, etc.
But only people closest to the application can make that assessment. Just offering it as an possible alternative to consider.