Specifications

DATA CENTER BEST PRACTICES
SAN Design and Best Practices 49 of 84
Here is a list of conguration differences between FICON and Open System environments that require VF LS
when mixing environments on the same switch or Director:
•The APTpolicy setting for FICON is not the same as Open Systems. FICON typically uses Port-Based Routing
(PBR) or Domain-Based Routing (DBR) with Lossless enabled. Open Systems uses Exchange-Based Routing
(EBR) without Lossless enabled.
•In-Order Delivery (IOD) is used in FICON environments. IOD is disabled in Open Systems environments.
•Security ACLs are required in cascaded FICON environments. Security ACLs are not used in Open Systems
environments.
•FICON Management Server (FMS) with Control Unit Port (CUP) is enabled only on FICON LSs.
•Brocade Network Advisor management of the LS in FICON mode vs. Open Systems mode.
Understand that using a VE_Port in a selected LS does not preclude that VE_Port from sharing an Ethernet
interface with other VE_Ports in other LSs. This is referred to as Ethernet Interface Sharing, refer to the
next section.
Ethernet Interface Sharing
An FCIP Trunk uses multiple Ethernet interfaces by assigning the circuits that belong to that trunk to different
Ethernet interfaces. Ipif are congured with IP addresses, subnet masks and an Ethernet interface, which
assigns the ipif to the interface. When the FCIP circuit is congured, the source IP address has to be one that
was used to congure an ipif, which in turn assigns the FCIP circuit to that Ethernet interface. It is possible to
assign multiple IP addresses and circuits to the same Ethernet interface by assigning multiple ipif to that same
interface, each with its own unique IP address.
Any one circuit cannot be shared across more than one Ethernet interface. An IP address/ipif/circuit can belong
only to one Ethernet interface. Thus, if more than one Ethernet interface is desired, you must use multiple
circuits. If the same IP address is attempted to be congured on more than one ipif, an error will occur, rejecting
the conguration.
It is possible to share an Ethernet interface with multiple circuits that belong to different VF LSs. The Ethernet
interface must be owned by the default switch (context 128). The ipif and iproute must also be congured within
the default switch. The VE_Port is assigned to the LS you want to extend with FCIP and is congured within that
LS. The FCIP tunnel is also congured within that LS using the IP addresses of the ipif that are in the default
switch. This permits efcient use of the 10 GbE interfaces.
Often, for purposes of redundancy and resiliency, an FCIP Trunk has circuits that extend out of both of the 10
GbE interfaces. Each 10 GbE interface (XGE) has “native” VE_Ports from one of the two groups (xge1:12-21
or xge0:22-31). If you wish to extend a circuit from VE_Port 12 through xge0, you must use something called
a cross-port. A cross-port requires an ipif and iproute that have been congured and explicitly designated for
cross-port use; otherwise, the circuit cannot be congured for the non-native 10 GbE interface. By merely
designating the ipif and iproutes to be used with non-native XGE interfaces, you can congure this type of circuit.
Figure 36 shows an example of two VF LS (50 and 60) on a Brocade FICON Director. The Ethernet interfaces are
in the default switch LS context 128. There are two circuits per FCIP Trunk. There is a red FCIP Trunk and a blue
FCIP Trunk. VE_Port 12 has one circuit that goes to its native xge1 interface and one circuit that goes to cross-
port xge0. VE_Port 22 has one circuit that goes to its native xge0 interface and one circuit that goes to cross-
port xge1. There are two 5 Gbps circuits emanating from each 10 GbE interface and transmitted over a physical
Ethernet link to an IP core gateway router within the data center.