NSX-T Load balancer Deployment Topology
Introduction to NSX-T Load balancer
Deployment Topology
In the NSX-T Data Center logical load balancer, the network traffic load is distributed among multiple servers to provide high-availability services to applications. Multiple servers are utilized to balance incoming service requests.
Using load balancing, you can map a virtual IP address to a group of pool servers. On the virtual IP address, the load balancer accepts TCP, UDP, HTTP, or HTTPS requests and decides which pool server to use.
NSX-T provides load balancing functionality, but it does so on Tier-1 gateway. Only T1 routers can support LB services, and each T1 gateway can support a single load balancer.
Load balancers can be deployed in either inline or one-arm mode.
One-Arm Topology
The load balancer is not in the path of traffic between the client and the server when in one-arm mode. Both the client and the server are mobile in this mode. In this situation, LB-SNAT is always necessary.
 |
| Fig 1.1- NSX-T One Armed Topology |
To guarantee that traffic from the server to the clients is routed to the load balancer, source NAT at LB is necessary. Therefore, unless we utilize XFF header for HTTP/HTTPS communication, the servers won't be able to view the clients' actual IP addresses.
Another topic of discussion is health check monitors, which are used to verify that each server is successfully executing. You may install health check monitors to examine the state of a server's health.
Inline Load Topology
The load balancer is in the route of traffic between the client and the server while in inline mode. The same tier-1 logical router cannot be used by clients and servers. In this instance, LB-SNAT is not necessary.
 |
| Fig 1.1- NSX-T Inline Topology |
In simple words, you can say that, The load balancer is in the route of traffic between the client and the server while in inline mode. If SNAT on the load balancer is not wanted, clients and servers shouldn't be linked to overlay segments on the same tier-1 logical router.
SNAT is necessary if clients and servers are linked to overlay segments on the same tier-1 logical router.
Other Articles you may interested in:
++++++++++++++++++++++++++++++++++++++++++++++++++++
VMware NSX-T 3.2 Updates and innovations - The Network DNA
Updates in VMware NSX-T 3.1.2 - The Network DNA
Unicast Traffic between VMs in VMware NSX-T environment - The Network DNA
Unicast Reverse Path Forwarding (uRPF) in VMware NSX-T - The Network DNA
Part 2: VMware NSX-T Replication Modes - The Network DNA
Part 1: VMware NSX Replication Modes - The Network DNA
Services High Availability in VMware NSX-T - The Network DNA
VMware NSX-T Logical Routing - The Network DNA
VMware NSX-T: Route Types on Tier-0 and Tier-1 Gateways - The Network DNA
Introduction to Micro segmentation in VMware NSX-T - The Network DNA
Dynamic Routing - BGP in VMware NSX-T - The Network DNA
NAT & DHCP Services on VMware NSX-T - The Network DNA
Gateway Firewall in VMware NSX-T - The Network DNA
VMware NSX-T Load Balancer, Virtual Server, Pool & Monitor Components - The Network DNA
NSX-T Controller Tables- TEP and ARP Tables - The Network DNA
Traffic Flooded in VMware NSX-T - The Network DNA
VMware NSX-T Data, Control & Management Plane - The Network DNA
VMware towards NGN : NSX-V and NSX-T differences - The Network DNA
VMware NSX-T Cluster types - The Network DNA
VMware NSX-T : N-VDS, teaming Policies and Uplink/pNIC - The Network DNA
NSX-T Terminologies: OVS, pNIC, vNIC, VNI & TEP - The Network DNA
Introduction to Transport zones in VMware NSX-T - The Network DNA
Functional components of VMWare NSX - The Network DNA
Introduction to VMware NSX for vSphere - The Network DNA
Basics about VMware NSX - The Network DNA
+++++++++++++++++++++++++++++++++++++++++++++++++++
