Sun Trunking Technology

Sun Trunking™ software provides the ability to aggregate multiple links between a pair of devices so that they work in parallel as if they were a single link. Once aggregated, these point-to-point links operate as a single highly available "fat pipe," providing increased network bandwidth as well as high availability. For a given link level connection, trunking enables you to add bandwidth up to the maximum number of network interface links supported.

Note

Sun Trunking is not included with the Solaris operating system. This is an unbundled software product.

Sun Trunking provides trunking support for the following network interface cards:

• Sun Quad FastEthernet adapter, qfe

• Sun GigabitEthernet adapter, ge

• Sun GigaSwift Ethernet UTP or MMF adapter, ce

• Sun Dual FastEthernet and Dual SCSI/P adapter, ce

• Sun Quad GigaSwift Ethernet adapter, ce

The key to enabling the trunking capability is the nettr command. This command can be used to trunk devices of the same technology together. Once trunked, a trunk head interface is established, and that interface is used by ifconfig to complete the configuration.

For example, if the two qfe instances (qfe0 and qfe1) need to be trunked, once the nettr command is complete, the trunk head would be assigned to qfe0. Then you could proceed to ifconfig to make the trunk operate under the TCP/IP protocol stack.

Trunking Configuration

The nettr(1M) utility is used to configure trunking. nettr(1M)can be used to:

• set up a trunk

• release a trunk

• display a trunk configuration

• display statistics of trunked interfaces

Following is the command syntax for nettr for setting up a trunk or modifying the configuration of the trunk members. The items in the square brackets are optional.

 nettr -setup head-instance device=<qfe | ce | ge>                             members=<instance,instance,.,.> [ policy=<number> ] 

Trunking Policies

MAC

• Is the default policy used by the Sun Trunking software. MAC is the preferred policy to use with switches. Most trunking-capable switches require use of the MAC hashing policy, but check your switch documentation.

• Uses the last three bits of the MAC address of both the source and destination. For two ports, the MAC address of the source and destination is first XORed: Result = 00, 01, which selects the port.

• Favors a large population of clients. For example, using MAC ensures that 50 percent of the client connections will go through one of two ports in a two-port trunk.

Round-Robin

• Is the preferred policy with a back-to-back connection used between the output of a transmitting device and the input of an associated receiving device.

• Uses each network interface of the trunk in turn as a method of distributing packets over the assigned number of trunking interfaces.

• Could have an impact on performance because the temporal ordering of packets is not observed.

IP Destination Address

• Uses the four bytes of the IP destination address to determine the transmission path.

• If a trunking interface host has one IP source address and it is necessary to communicate to multiple IP clients connected to the same router, then the IP Destination Address policy is the preferred policy to use.

IP Source Address/IP Destination Address

• Connects the source server to the destination based on where the connection originated or terminated.

• Uses the four bytes of the source and destination IP addresses to determine the transmission path.

• The primary use of the IP Source/IP Destination Address policy occurs where you use the IP virtual address feature to give multiple IP addresses to a single physical interface.

For example, you might have a cluster of servers providing network services in which each service is associated with a virtual IP address over a given interface. If a service associated with an interface fails, the virtual IP address migrates to a physical interface on a different machine in the cluster. In such an arrangement, the IP Source Address/IP Destination Address policy gives you a greater chance of using more different links within the trunk than would the IP Destination Address policy.