RtDb - Real Time Database
RtDb has an uncompromising real time ability to access 10,000,000 individual database records per second.
The organization of RtDb is comparable to a relational database. The data are kept in tables with rows and columns.
In a distributed system, each node may maintain a local copy of RtDb tables. In case of a change, the synchronization of these tables is performed according to priority through Multicast protocols.
Trigger mechanisms permit an application to define actions (installation of callback functions) that are activated automatically when data change in the RtDb.
Automatic synchronization of the RtDb with a relational database (ORACLE or SQL Server) or a file-persistence is possible.
RtIpC - Real Time Communication
Communication in the computer is performed via a "shared memory" and therefore makes possible local data transfer rates of more than 10,000,000 process data messages per second.
In the network, more than 50,000 (100 Mb LAN) process data messages per second may be sent. A specialty is the support of redundant networks to guarantee uninterrupted operation.
RtRM - Realtime Redundancy Manager
RtRM ensures the supervision of single processes and computers. This process controls the start of applications and, if a system failure occurs, also controls the computer switching for redundant systems ("failover").
The uninterrupted supervision of the crucial software functions is especially important for redundant systems as this supervision has to produce the criteria for computer switching.
Individual processes themselves can also recognize faults and report their occurrence to the RtRM.
RtRM also contains a watchdog function that supervises the processes. This means that they have to report configurable cycles at certain times. If they fail to do so, the processes themselves will be considered faulty. The individual computers supervise each other through cyclical messages on the redundant network.
The failure of a redundant computer causes a "failover". The redundancy may involve two or more computers. Times for timeout can be configured.
