Overview

A leading UK and Asia Public Transport Company are looking to expand their NoMachine Configuration and incorporating a new Linux Cluster. A Small number of 10-15 people would be connecting to this new Cluster but performing CPU intensive Modelling an Simulation.

Challenge

Supplying CPU intensive applications for Modelling and Simulation to multiple remote users was paramount.It was also mentioned as part of the specification that a Failover must be provided to improve resilience.

Results

WML found that the use of Enterprise Terminal Server combined with Enterprise Terminal Server Nodes provided the company with an option to produce their own custom Load balancing algorithms or from NoMachine’s predefined list.

Four general load balancing algorithms are available, two of which offer customisation of the algorithms in use:

Round-robin:

The server applies a plain round-robin algorithm to choose the node or a weighted round-robin algorithm if weight has been specified for each node. This is the default.

Custom:

The server uses the custom load-balancing algorithm specified in the NodeSelectionScript key.

Load-average:

The server chooses the node with the minimum load average provided by the system.

System-load:

The server selects the node with the lowest system load calculated according to the algorithm specified in the nxreportsystemload.pl script set on the node.

Each Enterprise Terminal Server would reside in different Data-Centers resulting in an active/passive pair for resillience and failover.