Load-Balancing

I was recently rebuilding part of my lab infrastructure, and as part of it, I wanted to migrate my vCD cells from two IPs each to a single IP (as this feature was added in vCD 8.10 ↗) for both the web UI and the console proxy. It simplifies provisioning, potential routing problems, and load-balancer configuration by having a single IP but separate ports for each service. Adding both services to a single IP is not new, Tomas Fjota wrote about it here ↗, however, there was some detail missing from his article to allow it to work behind a load-balancer....

While doing some research for NSX setups I found the urge to delve deeper into the calculations of some of ESXi’s load-balancing and teaming types that are available, below I have outlined the scenarios, calculations (where appropriate) and recommendations when it comes to choosing a NIC load balancing and teaming type. Virtual Port ID Your VMs all have single vNICs, You have multiple physical switches, the pNICs from the servers are striped across them, the switches aren’t stacked/don’t have an awareness of each other/are from different vendors (point here, completely different, no collaboration between equipment - any brownfield environment)....

Introduction I have been deploying a VDI solution recently based on the fantastic VMWare Horizon Suite ↗, one of the important points of deploying the Horizon View ↗ component of this is making it highly available and accessible from the outside for on-the-road users. The best way I have found to load-balance incoming connections (both internally and externally) is to set up a linux VM and run NginX ↗, which is a reverse caching proxy - it allows us to terminate the SSL connections and load-balance across our backend View Security Servers in a DMZ....