Data Interface Pairs

Topics | How To | Related Topics


Overview

Configure Data Interface Pairs

Sample Scenarios


Overview

A network interface name or address is the identifier by which a computer is known to a network. A computer can have multiple network interfaces when it has multiple Network Interface Cards (NIC) and each NIC may have unique interface names, such as amber1.company.com and amber2.company.com. These names can be used to recognize the same computer in two different domains (e.g., Local Area Network and Storage Area Network) or in the same domain. 

Using multiple NICs may reduce network congestion in situations where the software transfers a high volume of data using a network. This can be done as follows: 

Data Interface Pairs can be defined from the CommCell Console, or created in bulk with the DataInterfacePairConfig QScript in the command line interface. For information and instructions, see Command Line Interface - Qscripts. These Data Interface Pairs are then used to transfer data between the specified computers. Keep in mind that the data interface pairs are used to transfer data and control only between the specified pairs of computers. To communicate with all the other computers within the CommCell and the CommServe the default network interface is used. (The default network interface is also always used for conducting control communication with the CommServe computer.) You can also bind the services to a specific NIC as described in Binding Services to Specific Network Interface Cards (NIC).

The default network interface is used to both communicate and transfer data between Clients and MediaAgents, unless specific interface pairs are defined.

In situations where any one these interfaces fail, or is not functional, the software does not automatically switch over to the other network interface.

The software also prompts for the CommServe hostname during the MediaAgent and Client installation. In remote computers, it is important to specify either the name of NIC that is visible to the computer or the NIC that will be used for communication.


Configure Data Interface Pairs

Data Interface Pairs can be created between any two computers in a CommCell having multiple interfaces. This can be done in the following ways:

See Configure Data Interface Pairs for step-by step instructions.

Defining Data Interface Pairs when Multiple Interfaces have the Same Name

In this situation, it is necessary to instruct the CommCell components to use a specific Network Interface Card (NIC). You can do this using the following steps:

  1. Assign a unique name to the NIC that must be used, using the hosts file in the CommServe and all other components in the CommCell.
  2. Define the interface pairs that must be used between any two computers using the Data Interface Pairs wizard. You can define one interface pair between the computers. More than one interface pair is not supported.

Alternatively, you can type in the IP address of the specific NIC while defining the Data Interface Pairs.

Configuring Data Interface Pairs in CommNet

When you have multi-homed computers that has two or more network interface cards (NICs) in both the CommNet Server as well as the CommCell computer, you can configure the software to communicate across a specific NIC using the following procedures:


Sample Scenarios

Samples for Data Interface Pairs Usage

The following diagrams illustrate a few sample scenarios in which data interface pairs can be used:

In the above scenario all data is conducted through the backup domain back.com, thereby reducing network traffic on the production domains. The following data interface pairs have to be defined to accomplish this goal:

This interface pair can be used to conduct data over the backup domain, back.com. The default interface for the client is client.prod1.com, while the default network interface for MediaAgent1 is ma1.prod1.com.

This interface pair can be used to conduct auxiliary copy operations over the backup domain, back.com. The default network interface for the for MediaAgent2 is ma2.prod2.com

In the above scenario, the CommServe is located in one geographic domain, while the Client and MediaAgent are located in another different geographic domain. In such a situation, adding a third domain and defining the following pipeline pair between the Client and MediaAgent would result in efficient communication:

This interface pair can be used to conduct data over the backup domain, back.com. The default network interface for the client is client.prod.com and the default network interface for the MediaAgent is ma.prod2.com

In the above scenario, although all the components are in the same domain, (they could be in a different subnet) defining the following pipeline pair between Client2 and MediaAgent would result in better network communication:

Sample for Default Interface Name Usage

The following diagram illustrates a sample scenario in which defining the correct default interface name is beneficial.

In the above scenario, all the client computers are in the production domain while the MediaAgent and CommServe are attached to the backup domain back.com. By using the client1.back.com and client2.back.com as the default interface names for the clients, all the data protection operations and communication with the CommServe will be performed on the backup domain back.com.

Sample for CommServe Hostname Usage

The following diagram illustrates a sample scenario in which defining the correct CommServe hostname is beneficial.

In the above scenario, although the CommServe is installed with cs.net1.com as its default network interface, Client2 and MediaAgent2 must use the interface cs.net2.com.as the CommServe hostname.

Back To Top