Difference between real time system versus on line systems?

Answer 1

The prime (and possibly only significant) difference is in the kernel scheduler, which determines which processes get run, with what priority, and for how long.

In a typical "ordinary" OS, the scheduler is some sort of "fair-use" implementation, which insures that no one process monopolizes all the resources if other processes are waiting for CPU time. That is, this kind of scheduler is optimized for interactive use, and the ability to make sure all applications run in a reasonable amount of time.

In a RT OS, all applications are given strict priorities, and the scheduler is designed to give certain characteristics to each priority rating. For instance, it is entirely possible for a RT scheduler to allocate all resources to a single process, and not give anything to any other process, until that process is complete (or does something like block on I/O wait). The based concept behind a RT OS is that certain process priorities are more important than others, and this importance is paramount over any concept of "fair" resource use. In essence, a RT OS prioritizes work in such a manner that certain processes can preempt others and hog resources in ways that a typical OS would consider extremely "unfair". RT OSes are typically optimized for embedded controllers, which are designed to require certain actions be completed immediately, regardless of other demands on the system.

Answer 2

A "real time" database is a stand alone database. Therefore, all updates/deletes/changes are posted to the system immediately. A database resides on an operating system--a machine full of hardware that makes the database run. Disk drives, tape drives. So, here's how a database is built. (one being the first step and the bottom of the hierarchy). 1. Machine (this could be any Operating System: NT, Unix, Hp NonStop, Linux). The machine is also called a "node". The node will have a name. i.e.: LTSUnix2 2. A database has a name. i.e.: BBWSALES. Objects in this database are fully qualified--starting with the node and database names, followed by the drive name, then the file name. LTSUnix2\BBWSALES\drive3\wedsales.db1

In a distributed database, the database--BBWSALES, would be connected via database links to another database on another node. Views could be set up to pull data from both nodes/databases. Files could be ftp'd back and forth. Batch jobs could run and pull data from both places into a table.

Answer 3

Real time systems record transactions into the database immediately. One real time system are banks--you withdraw from your checking account via the ATM. Your checking account immediately records that transaction and lowers your balance. On line systems get data uploaded hourly, daily--whatever the business requirements are. Trending and analysis is usually done with this data.