The AIO instance (All-In-One) is the first choice for most projects for it's straight forward installation and configuration process. In this configuration, all the components and services of CYBERQUEST are residing on a single machine.
We recommend the AIO deployment for it's stability, ease of configuration and operation.
This version is most suitable for deployment in small to medium IT infrastructures.
By distributed edition (distributed architecture) we understand that CYBERQUEST can be deployed on multiple machines, in the same network or in different networks.
In the distributed architecture, CYBERQUEST can be deployed as AIO on multiple machines (for HA purposes etc.), or CYBERQUEST's component services are deployed on separate machines.
The main purpose for configuring CYBERQUEST in a distributed architecture is increase in performance by distributing load on multiple machines. Though it is very tempting for the consistent increase in performance, it is subject to increased service demands (due to the complexity of the instance) and third party software or hardware failures have a greater impact.
We recommend distributed architecture only for large projects, and we specifically design the solution for each such project.
Scaling Online DataStorage
This is done by deploying the Online DataStorage service on a separate machine. By usage of separate processing power for maintaining the NoSQL data repository, high ingestion (in EPS) and correlation values may be reached per CPU for the CYBERQUEST processing server.
For large amounts of data (above 5TB) an Online DataStorage cluster can be deployed. See chapter.DataStorage clustering.
Scaling Data Server
In some instances, where data must be collected from separate (and, in many cases, also remote) networks, the Data Server service (responsible with receiving the streams of data) can be deployed on separate machines. Also, the scaling of Data Server can be done for overall performance increase.
Scaling Rabbit MQ
By deploying the queueing service (powered by RabbitMQ) on separate machines the overall performance of the CYBERQUEST instance can be increased.
Scaling Data Storage (Archive)
In projects where data must be retained for extra long periods of time and with requirements of immediate access, the Data Storage service can be deployed on separate machines. This enables CYBERQUEST to maintain huge archive repositories. It also provides some increase in the overall performance of the CYBERQUEST instance.
Scaling Correlation Engine
Scaling the Correlation Engine, besides the Online DataStorage service, is the most performance increase scaling that can be implemented for a CYBERQUEST instance.