IBM Brocade, Инструкция по установке и эксплуатации

Страница: 24 / 74

12 Brocade 2-Port 10GbE Converged Network Adapter for IBM BladeCenter Installation and User’s Guide
Features, specifications, and standards
1
•
Interrupt moderation
Implements dynamic selection interrupt coalescing values based on traffic and system load
profiles. Traffic is continuously monitored to place in categories between “high throughput
sensitive” and “high latency sensitive.” Similarly, the host system is monitored regularly to
place it in categories between “highly loaded” and “minimally loaded.” The driver dynamically
selects interrupt coalescing values based on this profiling.
•
MSI-X
This is an extended version of Message Signaled Interrupts (MSI), defined in the PCI 3.0
specification. MSI-X helps improve overall system performance by contributing to lower
interrupt latency and improved utilization of the host CPU. MSI-X is supported by Linux RHEL5,
SLES 10 and 11, Windows 2008, and VMware ESX Server 3.5 and 4.0.
•
Network Priority
The Brocade 2-Port 10GbE Converged Network Adapter supports this feature, which provides a
mechanism to enable CEE flow control (802.1Qbb Priority-based Flow Control: Pause 802.1p)
on network traffic. In addition, it guarantees mutual exclusion of FCoE and network priorities to
ensure proper enhanced transmission selection (ETS). This feature is not supported on HBAs.
This feature does not need to be enabled on the adapter or switch. Specific CEE attributes,
including priorities for FCoE traffic, are configured on the FCoE switch. These attributes
propagate to the adapter CEE port through the DCBCXP. adapter firmware processes this
information and derives priorities for network traffic. The network driver is notified of the
network priority and tags both FCoE and Network frames with their priorities.
•
Priority-based flow control (802.1Qbb)
Defines eight priority levels to allow eight independent lossless virtual lanes. Pauses traffic
based on the priority levels and restarts traffic through a high-level pause algorithm.
•
Receive side scaling (RSS) feature for advanced link layer
Enables receive processing to be balanced across multiple processors while maintaining
in-order delivery of data, parallel execution, and dynamic load balancing.
•
TCP segmentation offload (TSO) and large send offload (LSO)
Large chunks of data must be segmented to smaller segments to pass through network
elements. LSO increases outbound throughput by reducing CPU overhead. Offloading to the
network card, where segmentation can be done by the Transmission Control Protocol (TCP), is
called TCP segmentation.
•
Virtual function-level statistics
•
VLAN (802.1Q)
A Virtual LAN (VLAN) is a way to provide segmentation of an Ethernet network. A VLAN is a
group of hosts with a common set of requirements that communicate as if they were attached
to the same LAN segment, regardless of their physical location. A VLAN has the same
attributes as a physical LAN, but it allows for end stations to be logically grouped together.