hub-tunnel scalability and limitations
search cancel

hub-tunnel scalability and limitations


Article ID: 218397


Updated On:


DX Unified Infrastructure Management (Nimsoft / UIM) Unified Infrastructure Management for Mainframe CA Unified Infrastructure Management SaaS (Nimsoft / UIM) CA Unified Infrastructure Management On-Premise (Nimsoft / UIM)


Customers need to understand Hub tunnel scalability, multi-tiering and tunnel limits.


- UIM 9.x or higher


Tunnels and Scalability 

Model to be used for greatest scalability -> Primary/core hub acting as a Tunnel client, secondary/client hubs as Tunnel Servers.

  • For Windows systems, sticking to less than 50 tunnel servers or less per each tunnel client due to the queue limit is a good practice.
  • The Primary Hub should be set up as the Tunnel 'client' in the majority of cases, especially for MSPs that need to deploy multiple hubs at various tiers. (Primary, Secondary, Tertiary etc.)
  • If the other Tunnel model is used where the Primary hub is acting as the Tunnel Server, and the secondary/remote hubs are acting as the Tunnel ‘clients,’ there is a limit of up to 48 SSL tunnels per single hub max but it can depend on other factors such as number of requests, traffic, bandwidth, hardware, network stability, or whether or not encryption is enabled.
  • Multi-processor hub with proper resources/recommended processor speed etc., 64-bit OS, etc., see installation guide for more details. Also, try to stick with a 3 GHz processor or higher.
  • All potential connectivity challenges such as local/remote firewall(s), DMZ/firewall rule misconfiguration, AV filtering/scanning etc. MUST be taken into consideration. Use a proxy server IF that is required for connectivity.
  • Linux tunnel server hubs are highly recommend and known to scale the best in terms of raw message rate and number of subscribers supported.
  • Note that in some cases tunnel client limits will end up much lower due to other factors but those potential causes/factors would need to be investigated by Support
Multi-Tiered Hubs

The following descriptions provide a generic Overview of hubs in a multi-tiered architecture layout. Note that Tier terminology may differ across environments but have the same general meaning or role.

Tier I

  • Primary Hub / Core Hub / NMS-Server / Master Hub / Monitoring Server
  • Example Tunnel Role: Tunnel Server


Tier II

  • Message Concentrator Hubs / Tunnel Hubs (dedicated), normally Linux to a.
  • Example Tunnel Role: Acts as Tunnel Client to Tunnel Server above it and as Tunnel Server to tunnel clients below it.


Tier III

  • Client/customer hubs / remote hubs / site hubs
  • Example Tunnel role: Tunnel client to Tunnel Server above it.

Tunnels between the Concentrator Hubs and the client/customer hubs/relay servers

  • Robots connected to their parent hubs.


Tip: For optimum scaling, the primary hub should be a Tunnel Client and scaled out across the other tiers in this fashion: Tunnel Client-> Tunnel Server/Client->Tunnel Servers. Note that a Hub tunnel can be configured as a hybrid (client and server).

Additional Information

Hub IM GUI Reference