Netgear AFT603 Product Data Sheet - Page 13

ProSAFE® LAN Access and Aggregation Chassis Switches Data Sheet M6100 series UDLD Support UDLD implementation detects unidirectional links physical ports (UDLD must be enabled on both sides of the link in order to detect an unidirectional link • UDLD protocol operates by exchanging packets containing information about neighboring devices • The purpose is to detect and avoid unidirectional link forwarding anomalies in a Layer 2 communication channel Both "normal-mode" and "aggressive-mode" are supported for perfect compatibility with other vendors implementations, including port "D-Disable triggering cases in both modes Datacenter Features (CLI only) The Priority Flow Control (PFC) is standardized by the IEEE 802.1Qbb specification and enables flow control per traffic class on IEEE 802 full-duplex links • By pausing congested priorities independently, highly loss sensitive protocols can share the same link with traffic that has different loss tolerances • The priorities are differentiated by the priority field of the 802.1Q VLAN header • PFC uses a new control packet defined in 802.1Qbb and therefore disables 802.3x standard flow control on PFC configured interfaces The Data Center Bridging Exchange Protocol (DCBX) is used by DCB devices to exchange configuration information with directly connected peers • The protocol is also used to detect misconfiguration of the peer DCB devices and optionally, for configuration of peer DCB devices • DCBX is deployed in support of lossless operation for FCoE or ISCSI traffic when all network elements in FCoE topologies are DCBX enabled • M6100 automatically detects if a peer is operating with either of the two DCBX versions (the CEE DCBX and the IEEE standard DCBX) by default • DCBX protocol supports the propagation of received configuration information for the following features: -- Enhanced Transmission Selection (ETS) -- Priority-based Flow Control (PFC) -- Application Priorities Enhanced Transmission Selection (ETS) provides an operational model for priority processing and bandwidth allocation for the switch in a DCB environment • Using priority-based processing and bandwidth allocations, different Traffic Class Groups (TCGs) within different types of traffic such as LAN, SAN and Management can be configured to provide bandwidth allocation or best effort transmit characteristics • CoS information is exchanged with peer DCBX devices using ETS TLVs • As part of the transmitted ETS TLVs, by default, DCBX advertises the following parameters on per port basis: -- Mapping between ingress ports 802.1p priority to Traffic Class Group (TCG) -- Bandwidth percentage (weight percentage) of each Traffic Class Group -- Scheduling algorithm for each Traffic Class Group • ETS TLVs are accepted from auto-upstream devices and propagated to auto-downstream devices • ETS may be configured on a port in manual mode and M6100 switch may become the source for ETS configuration in the network The FCoE Initialization Protocol (FIP) is used to perform the functions of FC_BB_E device discovery, initialization and maintenance • FIP uses a separate EtherType from FCoE to enable the distinction of discovery, initialization, and maintenance traffic from other FCoE traffic • FIP frames are standard Ethernet size (1518 Byte 802.1q frame) whereas FCoE frames are a maximum of 2240 bytes The FIP snooping capability is a frame inspection method used by FIP Snooping Bridges to monitor FIP frames and apply policies based upon the L2 header information in those frames, following the recommendations in Annex C of FC_BB_5 Rev 2.00 and supporting these features: • Auto-configuration of Ethernet ACLs based on information in the Ethernet headers of FIP frames • Emulation of FC point-to-point links within the DCB Ethernet network • Enhanced FCoE security/robustness by preventing FCoE MAC spoofing The FIP Snooping Bridge solution in M6100 supports the interior port role, the perimeter port role and the FCF-facing port role and is intended for use at the edge or the interior of the switched network • Perimeter or Edge port (connected directly to ENode) • FCF facing port (that receives traffic from FCFs targeted to the Enodes) Page 13 of 48