Fault tolerance technology provides a working environment even when problems occur. This is normally achieved through redundancy of system resources.

	Resource	Fault Tolerant Solution
	Files	Dfs Distributed File System
	Disks	RAID Redundant Array of Independent Disks
	Servers	MSCS Microsoft Cluster Server and NLB Network Load Balancing
	Power	UPS Uninterruptible Power Supply

Follow the links to the left to get details on Dfs, RAID, and Clusters .

NLB Network Load Balancing

Clients will connect to one of the servers in the NLB cluster but will not be aware of which server they connect to. Normally data is replicated between servers. If one server crashes, users will be serviced by another remaining server.

NLB is in Windows 2000 Advanced Server and DataCenter Server

• Supports up to 32 servers sharing one IP address
• designed for web farms

Internet
NLB	www.abc.com
www1.abc.com	www2.abc.com	www3.abc.com
çreplication è	çreplicationè

Combining NLB and MSCS

NLB and MSCS can be combined as in the following diagram. An example application would be an NLB web farm getting data from a SQL database MSCS Microsoft Cluster Server as shown below.

Front-end LAN					Internet
Network Load Balancing Hosts
Back-end LAN
Cluster Server with Shared storage