Hi,
We are planning of application migration to HANA. It is ticket handling system where HAN would be used to give real time monitoring/history data about activities.
We are considering option of ‘Letting HANA to do business logic of selecting next agent to handle ticket’.
(Bare example) Runtime data would be stored in Agent table with one related table HandledTicket (Agent can handle more tickets simultaneously). Each agent would have properties which would be criteria for ticket assignment. SELECT statement with complicated WHERE clause would be used to choose next agent (currently running on many instances of application server). State of the tickets/agents will be updated frequently, but those UPDATES does not have to be logged in transaction log neither persisted (There are separate History tables).
We would like to avoid performance loss caused by transaction, disk I/O. Runtime tables would be small, only active (currently handled) data will be stored in those tables. Maximum few (tens of) thousands agents and less than 100,000 tickets.
Is there possibility to create ‘In memory table’ (something like MS SQL server ‘table variable’) visible by multiple sessions?.
Or any other suggestion for high performance?
-Zdenko