A simple GRM L7 deployment is to use a DNS nameserver to direct client requests to one or more GRM L7 servers. This deployment is best for organizations for which a simple configuration is the highest priority. This type of deployment has the following disadvantages:
If all client requests are sent to a single GRM L7 server, the server can easily become overloaded when the load of client requests is high. To handle overloading, you can deploy multiple GRM L7 servers and allow the DNS server to return their IP addresses to clients in a round-robin fashion.
There is no way to guarantee that the GRM L7 server to which the DNS nameserver redirects requests is the closest one to that client. (The DNS nameserver uses the round-robin method of load balancing.)
For example, a DNS nameserver could send a request from a client in Australia to a GRM L7 server in the United States rather than to a closer GRM L7 server in Australia. In this case, the GRM L7 server in the United States would find the GRM agent cache closest to the client. However, sending a request to the GRM L7 server in the United States takes longer than sending a request to the GRM L7 server in Australia.
The following illustration shows a deployment with a DNS nameserver directing client requests to the GRM L7 server.
The following table describes the path of the request in the previous illustration.
Copyright © 1994-2002, Network Appliance, Inc. All rights
reserved.
Part No. 210-00374
File: grmdplm5.htm
Last updated: 11/12/02 17:45:11