Load Balancing, Server Monitoring and Successive server failover solutions

Geographic load balancing deployed entirely as a service has revolutionized the global load balancing industry. With no hardware to deploy and no software to configure, TZO-GEO offers a complete paradigm shift for true global traffic management.

With “N” number of participating server/s or IP addresses, as an inbound DNS query comes in to the TZO DNS infrastructure, we lookup the Source IP address of the DNS query. We then match that source IP to a database of IP addresses and geographic longitude and latitude definitions. When a geographic location is found for the source DNS query IP address, we then calculate which of the participating servers is closest in proximity to the originating DNS query. This all happens in milliseconds and poses no detectable latency to the standard DNS query.

As with all of our TZO-HA services, 2 geographically dispersed monitors are deployed per server or IP address that we are monitoring. This eliminates false positive notifications of any outages. Both monitors for a specific server need to report server unavailable for any action to take place. If an outage is detected the server or IP address is removed from the available list and immediate notifications are sent.

• Degraded Server performance

Also new with TZO-GEO is the ability to configure automatic actions upon degraded server performance. Inherent in our monitoring mechanism, is a round trip timer. We can establish a base line with this timer and therefore determine a point at which a server can be deemed in a degraded state. For example, if a normal average execution time for the monitor for a particular server is 100ms, we can configure the degraded state of that server to be defined when that execution time exceeds x. Our notification mechanism can react to this state change.

• Variable Distance Vector (VDV)

In addition to notification of this state change, it can also be configured so that the degraded server artificially appears farther away from inbound DNS requests. When a server reaches a degraded state, a “Degraded Scaling Factor” can be configured. The degraded scaling factor artificially “adds distance” to the known location of the server, thereby reducing the amount of traffic being sent to that server. This feature allows for servers that enter a degraded state to still participate in the load balancing scheme but in a reduced capacity. As the sever returns to a normal state, the degraded scaling factor is removed.