To state it clearly, SAP HANA is a relational database management system that has proved to be a path breaking development for SAP as a business. Like any DBMS it is adept in handling standard database operations, however, its greatest USP lies in its ability to perform advanced analytics and ETL capabilities. These two features are great assets for SAP as an ERP solution.
More and more clients are implementing HANA and it is becoming very important for ABAPers to write HANA compliant code to avoid rework.
So, what is changing from the programming perspective?
Before HANA, the approach was to avoid hitting the database multiple times if the same result can be achieved by one single query, even if it brings in some extra data. Idea was to bring the data to the application server from database and process it there. This import of data from the database server to application server was attributed to the 3-tier structure.
Now in HANA, the system provides a volatile memory and as a result, the query and processing happen at the same layer. As a result, no additional table buffering is needed. What has dramatically changed is the requirement of taking into account the cost of bringing data from the database into application service. Data processing statements can be directly utilized, something developers would avoid in a non-HANA system.
HANA is definitely the future for a while till machine learning takes away all the ABAP jobs. It might take some time, so, learning how to optimize/write ABAP code for HANA compliant systems can be key to success.