Distributed Processing

TOS Aurora runs on a cluster in a Kubernetes container orchestration environment. In it's simplest configuration, a single 'central' cluster is implemented on a single machine (node). If this primary data node doesn't have sufficient resources to run effectively or if you need to scale up to handle large workloads, you can implement a distributed architecture easily, using one or more of the following distributed processing mechanisms:

  • By adding worker nodes to the cluster, thus increasing the system's compute capability - see Multi-Node Cluster
  • By adding remote collectors - separate clusters in different geographic locations that interact with local devices and perform processing locally thus freeing up the central cluster - see Remote Collectors
  • By adding additional VIPs for receiving syslog traffic from devices- see Primary VIP Address

See also High Availability.