en  |  lt  |  de  |  ru
SoftWare | SoftWare  / Facebook SoftWare | SoftWare  / Twitter SoftWare | SoftWare  / Youtube


General principles of organisation and structure
Specialised server software implemented in the cluster architecture is used for communication between NT Adaptive controllers with the outside world and the solution of applied recognition tasks.
The use of the cluster architecture allows to achieve scalability and fault tolerance.
The logical foundation of the cluster is a virtual data bus constructed with the use of the message queue ideology (Message Queue – MQ). At a lower level, messages are transmitted via the protocol TCP/IP, which, in turn, can use the local network (possibly dedicated) built according to the technology 1G Ethernet or, if it necessary to increase the bandwidth and decrease latency, 10G/40G/100G Ethernet or InfiniBand.
Software cluster elements can be hosted on a single physical server. In this case, the cluster architecture is maintained and transport IPC can be used instead of TCP/IP to reduce the overhead costs of messaging.
Both servers and general-purpose computers, the characteristics of which (the number of cores, RAM capacity, etc.) are selected on the basis of the complexity of the tasks to be solved. Owing to the cluster architecture, the number and characteristics of the physical nodes may change over time and there is no requirement of using any identical hardware.

The software modules of the cluster can be conventionally divided into several categories, based on the list of tasks to be solved:

* Interface Modules
* Computing Modules
* Service Modules

Interface Modules

They receive data from external sources (MES systems, databases, etc.). On the basis of the received raw data, they form tasks for the Computing Modules and put them in a queue. They receive tasks from the queue for sending to the Adaptive Controllers NT Adaptive. At request, they form a visual representation of the current state of the cluster and its components (web interface, administrator command line, etc.).

Computing Modules

They extract data processing tasks from the queue and execute them (filtering, averaging, smoothing, format conversion, etc.).
On the basis of the obtained results, they create tasks and put them in the queue for other components of the cluster.

Service Modules

The maintenance of the infrastructure of the cluster (the current topology – lists of available resources)
Memorisation of the cluster operation results in a long-term storage while maintaining a time storage window
Logging of events in the cluster while maintaining a time window 
Reporting on the operation of the cluster with the required level of detail.


Neurotechnologies contacts
  • Lvovo str. 25, Vilnius, 09320, Lithuania
    Business centre 3BURĖS (3 Burės)
  • (+370) 606 21187
    (+370) 673 16915
  • info@neurotech.lt