Tutorial

Load Balancing Strategy

The load balancing provided by SuperMap iManager can compose a load balancing group of multiple GIS environments, improving the performance of GIS environment. The polling method, weighted polling method, IP Hash/weighted IP hash method, minimum connection method/Weighted minimum connection method are employed to bear load. The requests sent externally will be assigned to a GIS environment in the load balancing group evenly, and the requested server responds independently to the customer’s request. This allows you to acquire important data in a fast way and solve the high concurrency problem.

Polling/Weighted Polling

The principle of the polling method is that each time the requests from the user are assigned to the GIS environment in turn, starting from 1 until N (the number of GIS environments), and then restart the loop.

Weighted polling method assigns different weights to each server based on the polling method according to the different processing ability of the service. The number of connections received by a machine is assigned according to the weight. For example, you can set the processing capacity of the third machine is twice times the first machine, then the load balancer will allocate twice times the number of connections to the third machine.

Advantages: More adaptable. Independent of any information the clients , completely relies on the back-end server to make the choice. Assign client requests to each back-end server in a more reasonable way.

Disadvantages: Multiple requests from the same client may be assigned to different back-end servers for processing and cannot meet the need for session-keeping applications.

IP Hash/Weighted IP Hash

IP hash method assigns each request according to hash result, so that each visitor has a fixed back-end server to access.

Weighted IP hash method assigns different weights to each server based on the IP hash method according to the different processing ability of the service. The number of connections received by a machine is assigned according to the weight. For example, you can set the processing capacity of the third machine is twice times the first machine, then the load balancer will allocate twice times the number of connections to the third machine.

Advantages: It is better to allocate multiple requests from the same client to the same server, suitable for session-keeping needs: IP hash method preserves login information for the GIS environment.

Disadvantages: When there is a particular number of requests from an IP address at some point, the pressure on a back-end server may be very huge while other servers are idle.

Least Connections/Weighted Least Connections

The least connections method assigns new connections to a GIS environment with the least number of connections, which is very effective if there are various GIS environments with similar processing capability.

Weighted least connections method assigns different weights to each server based on the least connections method according to the different processing ability of the service. The number of connections received by a machine is assigned according to the weight. For example, you can set the processing capacity of the third machine is twice times the first machine, then the load balancer will allocate twice times the number of connections to the third machine.

Advantages: It is very effective if there are various GIS environments with similar processing capability.

Disadvantages: Schedule according to the current number of connections established in the GIS environment, while the resource consumption of the GIS environment in the process of processing the request is not considered. For the weighted least connections method, it can not adjust the weight of GIS environment in time according to the load condition during the work.