Failover for redundant Ethernet: The IBM 10/100 Ethernet adapter or the IBM 10/100 Etherjet™ PCI family of adapters are optional, redundant network interface card (NIC) adapters that you can install in your server. If you install this NIC adapter and connect it to the same logical segment as the primary Ethernet controller, you can configure the server to support a failover function. You can configure either the integrated Ethernet controller or the NIC adapter as the primary Ethernet controller. In failover mode, if the primary Ethernet controller detects a link failure, all Ethernet traffic that is associated with it is switched to the redundant (secondary) controller. This switching occurs without any user intervention. When the primary link is restored, the Ethernet traffic switches back to the primary Ethernet controller.

High Performance Ethernet Modes: Your Ethernet controller supports optional modes, such as teaming, priority packets, and virtual LANs, which provide higher performance and throughput for your server.

Teaming Mode: Your Ethernet controller provides options, called teaming options. These options increase throughput and fault tolerance when running with Windows NT 4.0 or NetWare 4.1x or later.

Adapter Fault Tolerance (AFT) provides automatic redundancy for your adapter. If the primary adapter fails, the secondary adapter takes over. Adapter fault tolerance supports from 2 to 4 adapters per team.

Adaptive Load Balancing (ALB) enables you to balance the transmission data flow among 2 to 4 adapters. ALB also includes the AFT option. You can use ALB with any 100BASE-TX switch.

Cisco Fast EtherChannel (FEC) creates a team of 2 to 4 adapters to increase transmission and reception throughput. FEC also includes the AFT option. You can only use FEC with a switch that has FEC capability.

Teaming requires you to install at least one additional Ethernet adapter. For additional information about the teaming modes, refer to the documentation that comes with these additional adapters.

Priority Packet Mode: Priority Packet is a traffic-prioritization utility that enables you to set up filters to process high-priority traffic before normal traffic. You can send information from critical nodes or applications with an indicated priority. Because you set this priority at the host or entry point of the network, the network devices can base forwarding decisions on priority information defined in the packet.

Priority Packet information is available on the IBM Networking Web site at

http://www.ibm.com/networking/support

Priority Packet prioritizes traffic based on priority filters. These are parameters you assign to outgoing (transmit) packets. Using the Priority Filter Wizard, you can set up predefined or custom priority filters based on a node medium access control (MAC) address, Ethernet type, or by various properties of the protocol and connector. Priority Packet provides two different methods for prioritizing traffic: IEEE 802.1p tagging and High Priority Queue.

IEEE 802.1p is a new IEEE standard for tagging, or adding additional bytes of information to packets with different priority levels. Packets are tagged with 4 additional bytes, which increase the packet size and indicate a priority level. When you send these packets out on the network, the higher priority packets are transferred first. Priority packet tagging (also known as Traffic Class Expediting) enables the adapter to work with other elements of the network (such as switches and routers) to deliver priority packets first. You can assign specific priority levels from 0 (low) to 7 (high).

You can assign values to packets based on their priority when you use the IEEE 802.1p standard for packet tagging. This method requires a network infrastructure that

66Hardware Maintenance Manual: xSeries 200

Page 74
Image 74
IBM x Series 200 manual Hardware Maintenance Manual xSeries