Abstract: Techniques are provided for two components in a distributed Fiber Channel Forwarder (FCF) to establish a link between them at a level of operation that is mutually compatible with respect to the various capabilities offered by the two components. A controlling switch in the distributed FCF may simultaneously operate at different levels with different Fiber Channel Data-Plane Forwarders (FDFs), on a per-pair basis. The level of operation is established at the granularity of an individual capability offered by a switching element. When switching elements are upgraded, the switching elements can dynamically switch to higher or lower levels of operation for any or all of the capabilities defined per pair of switching element.

Abstract: Techniques are provided for two components in a distributed Fiber Channel Forwarder (FCF) to establish a link between them at a level of operation that is mutually compatible with respect to the various capabilities offered by the two components. A controlling switch in the distributed FCF may simultaneously operate at different levels with different Fiber Channel Data-Plane Forwarders (FDFs), on a per-pair basis. The level of operation is established at the granularity of an individual capability offered by a switching element. When switching elements are upgraded, the switching elements can dynamically switch to higher or lower levels of operation for any or all of the capabilities defined per pair of switching element.

Abstract: Techniques are provided for synchronizing, in a distributed Fiber Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.

Abstract: Method to perform an operation comprising, receiving a login request from an endpoint connected to first physical port of a first switch module of a distributed network switch, wherein the distributed network switch comprises a plurality of switch modules, wherein each switch module comprises a plurality of ASICs, responsive to the login request, storing, in a FCDF database, an entry comprising an identifier of the first physical port and a unique identifier of the endpoint, and responsive to receiving, from a cFCF, a zoning update comprising the unique identifier of the endpoint and an FCID for the endpoint, storing the FCID in the FCDF database entry for the endpoint, identifying a first ASIC, of the plurality of ASICs of the first switch module, connected to the first physical port, and updating a zoning table of the first ASIC to include the zoning update.

Abstract: Provided are a method, a system, and a computer program product in which a plurality of switches are maintained in a cascaded configuration. A switch relays a switch fabric internal link services (SW-ILS) to generate a pair of exchanges comprising a first exchange and a second exchange at the switch. In response to a termination of the first exchange of the pair of exchanges, the second exchange of the pair of exchanges is terminated.

Abstract: Techniques are provided for synchronizing, in a distributed Fiber Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.

Abstract: Techniques are provided for synchronizing, in a distributed Fibre Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.

Abstract: Provided are computer program product, system, and method for determining whether a short or long message format is used to communicate zone information in a network. One large zone is implemented for end devices coupled to ports of data forwarders. A first formatted zoning distribution message is generated for the data forwarders identifying the data forwarder and ports on the data forwarders in the one large zone in response to determining that the data forwarder supports the short message format. A second formatted zoning distribution message is generated identifying the data forwarder, the ports on the data forwarders, and additional fields of information on the ports not included in the first formatted zoning distribution message in response to determining that the data forwarder does not support the short message format; and sending the generated first formatted or second formatted zoning distribution message to the data forwarder.

Abstract: A switch unit has one frame buffer pool for storing received frames and another frame buffer pool for storing large frames. The frame size in the large frame buffer pool may be optimized to the largest amount of data the switch unit that an FCoE switching is running on can support (i.e., a limitation of zone entries). Should free space be unavailable in the large frame buffer pool, or if a sequence grows bigger than can be supported, the switch unit may still continue to send response frames back to the sender. While the switch unit may store header information of the frame, the switch unit does not store the data of subsequent frames any longer. Once the sequence has been received completely, a rejection message is sent back with an appropriate error or reason code. The rejection message enables the sender to attempt a retransmission or cancel the current request altogether.

Abstract: A switch unit has one frame buffer pool for storing received frames and another frame buffer pool for storing large frames. The frame size in the large frame buffer pool may be optimized to the largest amount of data the switch unit that an FCoE switching is running on can support (i.e., a limitation of zone entries). Should free space be unavailable in the large frame buffer pool, or if a sequence grows bigger than can be supported, the switch unit may still continue to send response frames back to the sender. While the switch unit may store header information of the frame, the switch unit does not store the data of subsequent frames any longer. Once the sequence has been received completely, a rejection message is sent back with an appropriate error or reason code. The rejection message enables the sender to attempt a retransmission or cancel the current request altogether.

Abstract: Techniques are provided for distributing a fabric name to switching elements of a distributed switch configured to forward Fibre Channel over Ethernet (FCoE) frames. A distributed switch membership distributed (DMFD) message is transmitted that includes names of all switching elements in the distributed switch, and further includes a fabric descriptor that contains the fabric name.

Abstract: Techniques are provided for distributing a fabric name to switching elements of a distributed switch configured to forward Fibre Channel over Ethernet (FCoE) frames. A distributed switch membership distributed (DMFD) message is transmitted that includes names of all switching elements in the distributed switch, and further includes a fabric descriptor that contains the fabric name.

Abstract: Techniques are provided for synchronizing, in a distributed Fiber Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.

Abstract: Systems, methods, and computer program products to perform an operation comprising, receiving a login request from an endpoint connected to first physical port of a first switch module of a distributed network switch, wherein the distributed network switch comprises a plurality of switch modules, wherein each switch module comprises a plurality of ASICs, responsive to the login request, storing, in a FCDF database, an entry comprising an identifier of the first physical port and a unique identifier of the endpoint, and responsive to receiving, from a cFCF, a zoning update comprising the unique identifier of the endpoint and an FCID for the endpoint, storing the FCID in the FCDF database entry for the endpoint, identifying a first ASIC, of the plurality of ASICs of the first switch module, connected to the first physical port, and updating a zoning table of the first ASIC to include the zoning update.

Abstract: Systems, methods, and computer program products to perform an operation comprising, receiving a login request from an endpoint connected to first physical port of a first switch module of a distributed network switch, wherein the distributed network switch comprises a plurality of switch modules, wherein each switch module comprises a plurality of ASICs, responsive to the login request, storing, in a FCDF database, an entry comprising an identifier of the first physical port and a unique identifier of the endpoint, and responsive to receiving, from a cFCF, a zoning update comprising the unique identifier of the endpoint and an FCID for the endpoint, storing the FCID in the FCDF database entry for the endpoint, identifying a first ASIC, of the plurality of ASICs of the first switch module, connected to the first physical port, and updating a zoning table of the first ASIC to include the zoning update.

Abstract: Method to perform an operation comprising, receiving a login request from an endpoint connected to first physical port of a first switch module of a distributed network switch, wherein the distributed network switch comprises a plurality of switch modules, wherein each switch module comprises a plurality of ASICs, responsive to the login request, storing, in a FCDF database, an entry comprising an identifier of the first physical port and a unique identifier of the endpoint, and responsive to receiving, from a cFCF, a zoning update comprising the unique identifier of the endpoint and an FCID for the endpoint, storing the FCID in the FCDF database entry for the endpoint, identifying a first ASIC, of the plurality of ASICs of the first switch module, connected to the first physical port, and updating a zoning table of the first ASIC to include the zoning update.

Abstract: A switch unit has one frame buffer pool for storing received frames and another frame buffer pool for storing large frames. The frame size in the large frame buffer pool may be optimized to the largest amount of data the switch unit that an FCoE switching is running on can support (i.e., a limitation of zone entries). Should free space be unavailable in the large frame buffer pool, or if a sequence grows bigger than can be supported, the switch unit may still continue to send response frames back to the sender. While the switch unit may store header information of the frame, the switch unit does not store the data of subsequent frames any longer. Once the sequence has been received completely, a rejection message is sent back with an appropriate error or reason code. The rejection message enables the sender to attempt a retransmission or cancel the current request altogether.

Abstract: A switch unit has one frame buffer pool for storing received frames and another frame buffer pool for storing large frames. The frame size in the large frame buffer pool may be optimized to the largest amount of data the switch unit that an FCoE switching is running on can support (i.e., a limitation of zone entries). Should free space be unavailable in the large frame buffer pool, or if a sequence grows bigger than can be supported, the switch unit may still continue to send response frames back to the sender. While the switch unit may store header information of the frame, the switch unit does not store the data of subsequent frames any longer. Once the sequence has been received completely, a rejection message is sent back with an appropriate error or reason code. The rejection message enables the sender to attempt a retransmission or cancel the current request altogether.

Abstract: Techniques are provided for synchronizing, in a distributed Fibre Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.

Abstract: Techniques are provided for synchronizing, in a distributed Fibre Channel fabric or a distributed FCoE fabric in which FC frames are encapsulated in Ethernet frames, a controlling FCoE forwarder (cFCF) with the FCoE data-plane forwarder (FDF). The operation includes entering a recovery mode at the FDF. The FDF is modified based on a route distribution message provided by the cFCF which includes fabric-provided MAC addresses (FPMAs). The FDF notifies the cFCF of differences between the FPMAs supplied by the route distribution message with the FPMAs known by the FDF. The FDF leaves the recovery mode and sends an acknowledgement to the cFCF.