Global Server Load Balancing: Overview
AppNexus has three datacenters, one in Los Angeles, one in the New York Metro region, and one in Amsterdam, the Netherlands. DNS-based global server load balancing (also known as Global Load Balancing, Global Traffic Management, or GSLB) distributes traffic between identical applications at those datacenters. The two major benefits to this are:
- Reduced Latency. Having geographically disparate datacenters allows you and the outside world to reach your equipment quickly from anywhere on the globe. For example, users on the West coast of the United States are routed first to Los Angeles, and users on the East coast of the United States to New York. Users in Europe are routed to Amsterdam, the Netherlands.
- Availability. If an entire datacenter goes down, you maintain service. This is a distinct possibility; entire datacenters DO go down, because of ISP failures, multiple equipment failures, large scale power outages, or natural disasters.
How It Works
GSLB works at the DNS level to distribute requests. Each AppNexus datacenter has its own GSLB server, and all of the servers broadcast an identical IP address to the Internet. Any request for this IP address is automatically routed to the closest server via BGP. That server then chooses where to direct the user depending on the designated load-balancing ratio. Once a site has been chosen, the server responds to the user's request with the IP address of the local load-balancing pool at that site.
If one GSLB server fails, all traffic is directed to the other GSLB server. The GSLB servers also monitors the Local Load-Balancers to make sure they are up and running.
Example of GSLB Routing
To walk through an example of GSLB see our Global Load Balancing Documentation.
For information on configuring Global Load Balancing and setting load-balancing ratios, see Managing Global Server Load Balancing.
Next core concept: Storage