Open topic with navigation

Planning

Before you begin your Double-Take Availability installation, you should plan your implementation strategy. Ask yourself the following questions.

• What is the role of each server? Will this server be a source? Will this server be a target?
• Is the source server a Domain Controller? Or does it have another very specific role or configuration? You may want consider protecting the entire server in these cases.
• Is the source running Microsoft Exchange or Microsoft SQL? You should use the Double-Take Availability Application Manager to protect these applications.
• How much data will you be protecting? Can your target handle that amount of data?
• How much bandwidth is available between your source and target? Can your network handle the mirroring and replication traffic between the two servers? If the amount of change is greater than the bandwidth available, you may want to consider getting additional bandwidth or planning for disk queuing.

If there are concerns about resource utilization or how Double-Take Availability replication will impact the environment, you can profile the source server, the network links between the source and target, and the target server before installing Double-Take Availability to ensure that each component has adequate resources to handle the added load of replicating the data. Most environments do not require this type of analysis, but it may be needed if there are applications producing high-volume file writes or limited CPU, memory, disk, or network resources.

The best way to understand the impact of replication in an environment is to set up test equipment that simulates the production environment. However, if the resources to test in this manner are not available, resource utilization can be analyzed using Windows Performance Monitor and a utility to monitor network utilization. Performance data should be logged for a period that encompasses normal usage as well as any maintenance, backup, scheduled jobs, or batch processing that occurs. If utilization of any component is extremely high for a significant period of time, then it may be necessary to modify particular Double-Take Availability options. Keep in mind that some factors that are typically not in a test environment, such as backups and other applications using bandwidth, can affect resource utilization in the production environment.

After you have installed Double-Take Availability, you can do additional planning with the Throughput Diagnostic Utility, which determines your bandwidth needs by simulating a connection and recording data.

• Create a connection to the Diagnostics target. This connection uses a built-in null (non-existent) target to simulate a real connection. No data is actually transmitted across the network. Since there is no true connection, this connection type helps you plan your implementation strategy.
• After the connection has run for hours, or even days, use DTStat to convert the statistics file to a .csv file, which can be opened in a spreadsheet for analysis.

One method to avoid for planning purposes is estimating the amount of data that will be replicated in a given period using the amount of data backed up in a differential backup. Although this may be valid in some cases, it is usually not a good indicator because it is based on the differences in data at the time of backup. For example, if a 1 MB Microsoft Word document is saved ten times throughout the day, this will result in 10 MB of replication traffic because Word rewrites the entire file each time it is saved. However, this will only result in 1 MB being backed up for a differential backup.