8.3.17.0 - Force10 Networks
8.3.17.0 - Force10 Networks
8.3.17.0 - Force10 Networks
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
www.dell.com | support.dell.com<br />
Data center bridging satisfies the needs of the following types of data center traffic in a unified fabric:<br />
• LAN traffic consists of a large number of flows that are generally insensitive to latency requirements,<br />
while certain applications, such as streaming video, are more sensitive to latency. Ethernet functions as<br />
a best-effort network that may drop packets in case of network congestion. IP networks rely on<br />
transport protocols (for example, TCP) for reliable data transmission with the associated cost of greater<br />
processing overhead and performance impact.<br />
• Storage traffic based on Fibre Channel media uses the SCSI protocol for data transfer. This traffic<br />
typically consists of large data packets with a payload of 2K bytes that cannot recover from frame loss.<br />
To successfully transport storage traffic, data center Ethernet must provide no-drop service with<br />
lossless links.<br />
• Servers use InterProcess Communication (IPC) traffic within high-performance computing clusters to<br />
share information. Server traffic is extremely sensitive to latency requirements.<br />
To ensure lossless delivery and latency-sensitive scheduling of storage and service traffic and I/O<br />
convergence of LAN, storage, and server traffic over a unified fabric, IEEE data center bridging adds the<br />
following extensions to a classical Ethernet network:<br />
• 802.1Qbb - Priority-based Flow Control (PFC)<br />
• 802.1Qaz - Enhanced Transmission Selection (ETS)<br />
• 802.1Qau - Congestion Notification<br />
• Data Center Bridging Exchange (DCBX) protocol<br />
Note: In FTOS version <strong>8.3.17.0</strong>, only the PFC, ETS, and DCBX features are supported in data center<br />
bridging.<br />
Priority-Based Flow Control<br />
In a data center network, priority-based flow control (PFC) manages large bursts of one traffic type in<br />
multiprotocol links so that it does not affect other traffic types and no frames are lost due to congestion.<br />
When PFC detects congestion on a queue for a specified priority, it sends a pause frame for the 802.1p<br />
priority traffic to the transmitting device. In this way, PFC ensures that large amounts of queued LAN<br />
traffic do not cause storage traffic to be dropped, and that storage traffic does not result in high latency for<br />
high-performance computing (HPC) traffic between servers.<br />
PFC enhances the existing 802.3x pause and 802.1p priority capabilities to enable flow control based on<br />
802.1p priorities (classes of service). Instead of stopping all traffic on a link (as performed by the<br />
traditional Ethernet pause mechanism), PFC pauses traffic on a link according to the 802.1p priority set on<br />
a traffic type. You can create lossless flows for storage and server traffic while allowing for loss in case of<br />
LAN traffic congestion on the same physical interface.<br />
Figure 6-1 shows how PFC handles traffic congestion by pausing the transmission of incoming traffic with<br />
dot1p priority 3.<br />
52 | Data Center Bridging (DCB)