10
Copyright © 2000 - 2005 KEMP Technologies, Inc. All Rights Reserved.
High availability of services and applications – since services are replicated across multiple machines
within the farm, the loss of a single server does not result in a total loss of service for the customer –
worst case, a drop in performance may be experienced.
Greater flexibility – server machines can be added and removed from the farm on demand resulting in
an immediate effect. Services – sometimes referred to as Virtual Services in the context of Load
Balancers – can be trivially tuned and shuffled around the server farm, thus ensuring optimal usage of
resources.
Improved performance – intelligent load balancing algorithms ensure that requests are directed to
those machines in the server farm that can most effectively handle them.
The LoadMaster Load Balancer provides a more reliable, flexible and cost effective solution to addressing high
traffic load and mission critical applications, resulting in a high Quality of Service (QoS) at a low Total Cost of
Ownership (TCO).
2. Considerations in Getting Started
Skip this section if you are already familiar with the basic Load Balancer functions and know how to set up a
Balancer-Server combination.
The Installation and Configuration Guide describes how to go about installing and configuring your LoadMaster
such that load-balanced services (virtual services) can be supported. However, before doing so, the following
considerations covered by the LoadMaster documentation should be taken into account when setting up your
LoadMaster for the first time:
What sort of LoadMaster network topology best suits my application? [See section C of this guide]
Do my real servers require publicly routable IP addresses or can they be “hidden” behind the LoadMaster on a
private network segment? [See section C of this guide]
Does my Application require the Balancer and Real Servers to be part of a flat-based topology? [See section C
of this guide]
Will network connections be initiated from within the application farm to the outside, as well as from the
outside to within the farm? [See section D of this guide]
Do I require a High Availability support in the form of an active/hot-standby redundant cluster? [See section E
of this guide]
How do I intend to replicate my application across multiple real server machines?
What kind of Virtual Service settings best suits my application? [See section F of this guide]
Will a round robin balancing algorithm do or do I need to take aspects of my application into consideration and
report these back to the Balancer? [See section F of this guide]
Is connection persistency at all an issue for my application? [See section G and H of this guide]
Will source-based IP persistency suffice or do I need to take layer 7 aspects into consideration? [See section G
and H of this guide]
Do I wish to integrate the LoadMaster into my current SNMP environment? [See section
J of this guide]
Which forms of real server health checking will best suite my application? [See section K of this guide]
