Integrated fibre channel support in an ethernet fabric switch

Abstract

One embodiment of the present invention provides a switch. The switch includes a packet processor and a device management module. During operation, the packet processor constructs a notification message containing a Transparent Interconnection of Lots of Links (TRILL) Routing Bridge (RBridge) identifier associated with a Fiber Channel router. The device management module operates in conjunction with the packet processor and terminates TRILL forwarding for a received TRILL packet with the RBridge identifier as an egress RBridge identifier.

Description

BACKGROUND

Field

The present disclosure relates to network management. More specifically, the present disclosure relates to a method and system for efficient integration of a storage area network to a Transparent Interconnection of Lots of Links (TRILL) network.

Related Art

The exponential growth of the Internet has made it a popular delivery medium for multimedia applications, such as video on demand and television. Such applications have brought with them an increasing demand for bandwidth. As a result, equipment vendors race to build larger and faster switches with versatile capabilities, such as Fibre Channel (FC) connectivity, to move more traffic efficiently. However, the size of a switch cannot grow infinitely. It is limited by physical space, power consumption, and design complexity, to name a few factors. Furthermore, switches with higher capability are usually more complex and expensive. More importantly, because an overly large and complex system often does not provide economy of scale, simply increasing the size and capability of a switch may prove economically unviable due to the increased per-port cost.

A flexible way to improve the scalability of a switch system is to build a fabric switch. A fabric switch is a collection of individual member switches. These member switches form a single, logical switch that can have an arbitrary number of ports and an arbitrary topology. As demands grow, customers can adopt a “pay as you grow” approach to scale up the capacity of the fabric switch.

Meanwhile, layer-2 (e.g., Ethernet) switching technologies continue to evolve. More routing-like functionalities, which have traditionally been the characteristics of layer-3 (e.g., Internet Protocol or IP) networks, are migrating into layer-2. Notably, the recent development of the Transparent Interconnection of Lots of Links (TRILL) protocol allows Ethernet switches to function more like routing devices. TRILL overcomes the inherent inefficiency of the conventional spanning tree protocol, which forces layer-2 switches to be coupled in a logical spanning-tree topology to avoid looping. TRILL allows routing bridges (RBridges) to be coupled in an arbitrary topology without the risk of looping by implementing routing functions in switches and including a hop count in the TRILL header.

As more data-intensive applications are deployed, high-performance layer-2 communication to a storage area network is becoming progressively more important as a value proposition for network architects. It is desirable to provide data flows from a storage device with a quality of service while providing the flexibility and ease of deployment of the layer-2 communication.

While TRILL brings many desirable features to a network, some issues remain unsolved in efficient Fibre Channel integration.

SUMMARY

One embodiment of the present invention provides a switch. The switch includes one or more ports capable of receiving packets based on a first protocol, a packet processor, and a device management module. During operation, the packet processor constructs a notification message containing a switch identifier which is based on a second protocol and associated with a router that operates on the first protocol. The device management module operates in conjunction with the packet processor and terminates forwarding for a packet which is encapsulated with the second protocol and has the switch identifier based on the second protocol as an egress RBridge identifier.

In a variation on this embodiment, the first protocol is a Fibre Channel protocol. The second protocol is a Transparent Interconnection of Lots of Links (TRILL) protocol. The switch identifier is a Routing Bridge (RBridge) identifier. The router that operates on the first protocol is a Fibre Channel router. Furthermore, the packet processor extracts a Fibre Channel packet from a received TRILL packet. The destination address of the Fibre Channel packet corresponds to the Fibre Channel router.

In a further variation on this embodiment, the Fibre Channel packet is encapsulated in an Ethernet frame.

In a variation on this embodiment, the device management module identifies a media access control (MAC) address of the router that operates on the first protocol as a local MAC address associated with the switch.

In a variation on this embodiment, the switch also includes an identification module and a flow control module. The identification module identifies a class of service associated with the Fibre Channel router. The flow control module configures priority-based flow control associated with the class of service.

In a further variation on this embodiment, the packet processor creates a notification message for a remote switch in response to identifying the class of service. The notification message contains the class of service associated with the priority-based flow control configured for the switch.

In a variation on this embodiment, the packet processor is further configurable to construct a notification message for a remote switch, wherein the notification message contains information learned locally.

In a variation on this embodiment, the switch also includes a forwarding module which makes local routing decision for a TRILL packet using a routing protocol instance. The TRILL packet contains a Fibre Channel packet or a non-Fibre Channel packet.

In a further variation on this embodiment, the routing protocol corresponds to one of: 1) Open Shortest Path First (OSPF); 2) Intermediate System To Intermediate System (IS-IS); and 3) Distance-vector.

In a variation on this embodiment, the switch also includes a fabric switch management module which maintains a membership in a fabric switch. Such a fabric switch accommodates a plurality of switches and operates as a single logical switch.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1A illustrates an exemplary TRILL network coupled to virtualized Fiber Channel (FC) routers, in accordance with an embodiment of the present invention.

FIG. 1B illustrates an exemplary TRILL network coupled to virtualized FC routers with hierarchical interconnections, in accordance with an embodiment of the present invention.

FIG. 2A illustrates an exemplary network configuration of a TRILL network coupled to virtualized FC routers, in accordance with an embodiment of the present invention.

FIG. 2B illustrates an exemplary network configuration of a TRILL network with hybrid RBridges coupled to servers and virtualized FC routers, in accordance with an embodiment of the present invention.

FIG. 3A presents a flowchart illustrating the process of a hybrid RBridge distributing priority-based flow control configuration information to other RBridges in a TRILL network, in accordance with an embodiment of the present invention.

FIG. 3B presents a flowchart illustrating the process of an RBridge self-configuring priority-based flow control based on received information from another RBridge in a TRILL network, in accordance with an embodiment of the present invention.

FIG. 4A presents a flowchart illustrating the process of a hybrid RBridge distributing virtualized FC connectivity information to other RBridges in a TRILL network, in accordance with an embodiment of the present invention.

FIG. 4B presents a flowchart illustrating the process of an RBridge learning about other networking devices via a TRILL port, in accordance with an embodiment of the present invention.

FIG. 5 presents a flowchart illustrating the process of an RBridge forwarding an Ethernet-encapsulated FC packet received from an end device, in accordance with an embodiment of the present invention.

FIG. 6 presents a flowchart illustrating the process of a hybrid RBridge processing a TRILL packet, in accordance with an embodiment of the present invention.

FIG. 7 illustrates an exemplary RBridge, in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION

The following description is presented to enable any person skilled in the art to make and use the invention, and is provided in the context of a particular application and its requirements. Various modifications to the disclosed embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the present invention. Thus, the present invention is not limited to the embodiments shown, but is to be accorded the widest scope consistent with the claims.

Overview

In embodiments of the present invention, the problem of extending a Fibre Channel (FC) fabric domain to a Transparent Interconnection of Lots of Links (TRILL) network is solved by facilitating packet forwarding in the TRILL network toward FC routers. Typically, a storage area network is dedicated to provide access to unit-level data (e.g., a data block) from target storage devices. Such a storage area network can make a storage device, such as a disk array, accessible to a remote server (i.e., a server not locally coupled to the storage device) such a way that the target appears as a storage device locally coupled to the server. Fibre Channel (FC) is a networking technology designed for efficiently move data through a storage area network. More details on FC network architecture, protocols, naming/address conventions, and various standards are available in the documentation available from the NCITS/ANSI T11 committee (www.t11.org) and publicly available literature, such as “Designing Storage Area Networks,” by Tom Clark, 2nd Ed., Addison Wesley, 2003, the disclosures of which are incorporated by reference in their entirety herein.

FC provides quality of service to a data flow between a server and a storage device, such as a RAID array or a tape backup device. However, to benefit from FC, a respective server requires a dedicated FC network interface card. Because a server typically exchanges information with a non-FC device via a layer-2 communication protocol, such as Ethernet, the server needs another network interface card (e.g., Ethernet network interface card) to communicate with the non-FC device. As a result, the server requires at least two interface cards (i.e., an Ethernet and an FC interface cards) to operate. Equipping a respective server with multiple interface cards in a large server farm incurs a high deployment cost. More importantly, multiple interfaces cards in a respective server lead to substantial power consumption, a high operational cost, and a large carbon foot-print for the server farm.

To solve this problem, FC routing capability can be integrated with an RBridge. An RBridge with FC routing capability can be referred to as a hybrid RBridge. One or more hybrid RBridges can be deployed in a TRILL network to couple a storage area network (i.e., an FC fabric) with the TRILL network. A respective hybrid RBridge presents the locally coupled FC routers as virtual RBridges to rest of the TRILL network and facilitate packet forwarding towards the FC routers in the TRILL network. Note that the rest of the TRILL network perceives a respective virtual RBridge as a regular RBridge coupled to the hybrid RBridge. A layer-2 (e.g., Ethernet) interface card couples a respective server with the TRILL network. To communicate with a target storage device, the server creates an FC packet for a corresponding FC router, encapsulates the FC packet in an Ethernet frame, and forwards the frame via the Ethernet interface card to the TRILL network. This encapsulation can be referred to as Fibre Channel over Ethernet (FCoE). An FC packet encapsulated in an Ethernet frame can be referred to as an FCoE frame.

Upon receiving the FCoE frame, an ingress RBridge identifies the destination of the frame to be a virtual RBridge. The ingress RBridge in the TRILL network is not aware of the FC routing capability of the hybrid RBridge. Hence, the ingress RBridge encapsulates the frame in a TRILL packet and forwards the frame toward the virtual RBridge. When the hybrid RBridge associated with the virtual RBridge receives the packet, the hybrid RBridge extracts the FC packet and forwards the FC packet to the FC router corresponding to the virtual RBridge. The FC router, in turn, forwards the frame to the target. In this way, the hybrid RBridge enables the server to communicate with a target storage device with the flexibility of a TRILL network and the performance of an FC fabric while using a single interface.

In some embodiments, the TRILL network is a fabric switch and a respective RBridge in the TRILL network is a member switch of the fabric switch. The fabric switch can be an Ethernet fabric switch. In an Ethernet fabric switch, any number of switches coupled in an arbitrary topology may logically operate as a single switch. Any new switch may join or leave the fabric switch in “plug-and-play” mode without any manual configuration. A fabric switch appears as a single logical switch to the end device.

In a fabric switch, the hybrid RBridge also virtualizes a respective FC router and presents the FC router to the rest of the fabric switch as a virtual member switch coupled to the hybrid RBridge. In some embodiments, the control plane of the fabric switch operates as a Fibre Channel fabric. Consequently, the hybrid RBridge can use the domain identifier of an FC router as the identifier to the virtual member switch. In this way, the hybrid RBridge extends the FC fabric domain to the fabric switch domain.

Although the present disclosure is presented using examples based on TRILL and FC, embodiments of the present invention are not limited to networks defined using TRILL, FC, or a particular Open System Interconnection Reference Model (OSI reference model) layer. For example, embodiments of the present invention can also be applied to a multi-protocol label switching (MPLS) network. In this disclosure, the term “TRILL network” is used in a generic sense, and can refer to any networking layer, sub-layer, or a combination of networking layers. The term “Fibre Channel” or “FC” is also used in a generic sense, and can refer to any networking technology for a storage area network.

The term “RBridge” refers to routing bridges, which are bridges implementing the TRILL protocol as described in Internet Engineering Task Force (IETF) Request for Comments (RFC) “Routing Bridges (RBridges): Base Protocol Specification,” available at http://tools.ietf.org/html/rfc6325, which is incorporated by reference herein. Embodiments of the present invention are not limited to application among RBridges. Other types of switches, routers, and forwarders can also be used.

In this disclosure, the term “edge port” refers to a port on an RBridge which sends/receives data frames in native Ethernet format. The term “TRILL port” refers to a port which sends/receives data frames encapsulated with a TRILL header and outer MAC header.

The term “end device” can refer a host, a conventional layer-2 switch, or any other type of network device. Additionally, an end device can be coupled to other switches or hosts further away from a network. An end device can also be an aggregation point for a number of network devices to enter the network. From a storage area network's perspective, an end device can be a server requesting data from the storage area network. In this disclosure, the terms “end device” and “sever” are used interchangeably.

The term “switch identifier” refers to a group of bits that can be used to identify a switch. If the switch is an RBridge, the switch identifier can be an “RBridge identifier.” If the switch is an FC router, the switch identifier can be a “domain identifier.” Note that the TRILL standard uses “RBridge ID” to denote a 48-bit Intermediate-System-to-Intermediate-System (IS-IS) ID assigned to an RBridge, and “RBridge nickname” to denote a 16-bit value that serves as an abbreviation for the “RBridge ID.” In this disclosure, “switch identifier” is used as a generic term, is not limited to any bit format, and can refer to any format that can identify a switch. The term “RBridge identifier” is used in a generic sense, is not limited to any bit format, and can refer to “RBridge ID,” “RBridge nickname,” or any other format that can identify an RBridge. The term “domain identifier” is also used in a generic sense, is not limited to any bit format, and can refer to any group of bits that uniquely identify a switch in an FC fabric.

The term “frame” refers to a group of bits that can be transported together across a network. “Frame” should not be interpreted as limiting embodiments of the present invention to layer-2 networks. “Frame” can be replaced by other terminologies referring to a group of bits, such as “packet,” “cell,” or “datagram.”

The term “switch” is used in a generic sense, and it can refer to any standalone or fabric switch operating in any network layer. “Switch” should not be interpreted as limiting embodiments of the present invention to layer-2 networks. Any device that can forward traffic to an end device can be referred to as a “switch.” Examples of a “switch” include, but are not limited to, a layer-2 switch, a layer-3 router, a TRILL RBridge, an FC router, or an FC switch.

The term “Ethernet fabric switch” refers to a number of interconnected physical switches which form a single, scalable logical switch. In a fabric switch, any number of switches can be connected in an arbitrary topology, and the entire group of switches functions together as one single, logical switch. This feature makes it possible to use many smaller, inexpensive switches to construct a large fabric switch, which can be viewed as a single logical switch externally.

Network Architecture

FIG. 1A illustrates an exemplary TRILL network coupled to virtualized FC routers, in accordance with an embodiment of the present invention. As illustrated in FIG. 1, a TRILL network 100 includes RBridges 101, 102, 103, 104, 105, 106, and 107. End devices 112 and 114 are servers coupled to network 100 via RBridges 104 and 103, respectively. RBridges in network 100 use edge ports to communicate to end devices and TRILL ports to communicate to other RBridges. Data frames transmitted and received via TRILL ports are encapsulated in TRILL headers. For example, RBridge 104 is coupled to end device 112 via an edge port and to RBridges 105, 101, and 102 via TRILL ports. In this example, RBridges 101, 102, and 103 are hybrid RBridges with FC routing capabilities, and coupled to storage area network 150.

In some embodiments, storage area network 150 is an FC fabric and includes FC routers 122, 124, and 126. Target storage devices 142, 144, and 146 are coupled to FC routers 122, 124, and 126, respectively. FC fabric 150 is dedicated to provide access to data blocks from targets 142, 144, and 146. Typically, to access targets 142, 144, and 146, servers 112 and 114 require dedicated FC network interface cards. Servers 112 and 114 also need another network interface card (e.g., Ethernet network interface card) to communicate with non-FC devices. Equipping server 112 and 114 with multiple interface cards incurs a high deployment cost and substantial power consumption.

To solve this problem, hybrid RBridges 101, 102, and 103 present FC routers 122, 124, and 126 as virtual RBridges 132, 134, and 136, respectively, to RBridges 104, 105, 106, and 107. Note that RBridges 104, 105, 106, and 107 perceive virtual RBridges 132, 134, and 136 as regular RBridges coupled to hybrid RBridges 101, 102, and 103, respectively. In some embodiments, network 100 is a fabric switch and the RBridges in network 100 are member switches of the fabric switch. This fabric switch appears as a single logical switch to servers 112 and 114.

In some embodiments, the control plane for network 100 operates as a FC fabric and RBridge identifiers in network 100 are in the same format as the domain identifier of FC routers 122, 124, and 126. Hybrid RBridges 101, 102, and 103 advertize the domain identifiers of FC routers 122, 124, and 126 as the identifiers of virtual RBridges 132, 134, and 136. In this way, hybrid RBridges 101, 102, and 103 extend the domain of storage area network 150 to the domain of network 100. As a result, a single routing protocol instance in a respective RBridge in network 100 can make routing decisions for TRILL packets containing Fibre Channel or non-Fibre Channel packets. Examples of the routing protocol can include, but are not limited to Open Shortest Path First (OSPF) routing protocol, Intermediate System To Intermediate System (IS-IS) routing protocol, and Distance-vector routing protocol.

During operation, hybrid RBridges 101, 102, and 103, optionally, learns about targets 142, 144, and 146, from FC routers 122, 124, and 126, respectively. Hybrid RBridges 101, 102, and 103, in turn, create and send notification messages to RBridges 104, 105, 106, and 107 indicating that targets 142, 144, and 146 are coupled to virtual RBridges 132, 134, and 136, respectively. To communicate with target 142, server 112 creates an FC packet, encapsulates the FC packet in an Ethernet frame, and forwards the FCoE frame to ingress RBridge 104. While constructing the FCoE frame, server 112 assigns a domain identifier of FC router 122 as the destination address of the inner FC packet and the MAC address of FC router 122 as the destination address of the Ethernet frame.

In some embodiments, upon receiving the FCoE frame, RBridge 104 checks the encapsulated FC packet and identifies destination target 142 to be coupled to virtual RBridge 132. RBridge 104 is not aware of the FC routing capability of hybrid RBridge 101. Hence, RBridge 104 simply uses regular TRILL routing for forwarding the FCoE frame to virtual RBridge 132. Using the local routing protocol instance, RBridge 104 determines the next-hop RBridge to be RBridge 101. RBridge 104 then encapsulates the FCoE frame in a TRILL packet destined to virtual RBridge 132 (i.e., with the identifier of virtual RBridge 132 as the egress RBridge identifier) and forwards the packet to hybrid RBridge 101. In some embodiments, hybrid RBridge 101 marks the RBridge identifier of virtual RBridge 132 as a local RBridge identifier (i.e., as belonging to hybrid RBridge 101). Upon receiving the TRILL packet, hybrid RBridge 101 recognizes the RBridge identifier as a local RBridge identifier and terminates TRILL forwarding for the packet.

In some embodiments, hybrid RBridge 101 marks the MAC address of FC router 122 as a local MAC address. The MAC address of FC router 122 can be derived from the domain identifier of FC router 122. Hybrid RBridge 101 then extracts the FCoE frame from the TRILL packet and recognizes the destination address of the FCoE frame to be a local MAC address. In other words, hybrid RBridge 101 considers the FCoE frame to be destined to itself. As a result, hybrid RBridge 101 decapsulates the Ethernet frame, promotes the internal FC packet to the upper layer, and forwards the FC packet to FC router 122 coupled to target 142. Similarly, server 114 can communicate with target 144 via hybrid RBridges 103 and 102, and FC router 124. In this way, hybrid RBridge 101, 102, and 103 allow servers 112 and 114 to access targets 142, 144, and 146 via network 100 using one layer-2 interface.

In some embodiments, network 100 ensures automatic priority-based flow control configuration for all RBridges in network 100. Upon identifying FC router 122, hybrid RBridge 101 obtains the class of service associated with the FC traffic from FC router 122 and configures priority-based flow control for the class of service. In the same way, hybrid RBridges 102 and 103 also obtains the class of service for FC traffic from FC routers 124 and 126, respectively. Hybrid RBridges 101, 102, and 103 propagate the associated configuration information to respective other RBridges in network 100. Upon receiving the configuration information, other RBridges self-configure the priority-based flow control for that specific class of service. In this way, RBridges in network 100 provide a desired quality of service to the FCoE frames.

FIG. 1B illustrates an exemplary TRILL network coupled to virtualized FC routers hierarchical interconnections, in accordance with an embodiment of the present invention. Components in FIG. 1B are the same as in FIG. 1A, so the same numerals are used to denote them. In this example, FC router 128 is coupled to FC routers 122, 124, and 126. FC router 128 virtualizes FC routers 122, 124, and 126, and presents targets 142, 144, and 146 to hybrid RBridges 101, 102, and 103 as coupled to FC router 128. Hybrid RBridges 101, 102, and 103, in turn, present FC router 128 as virtual RBridge 138 to RBridges 104, 105, 106, and 107. To ensure a separation between network 100 and FC fabric 150, a respective local FC port of hybrid RBridges 101, 102, and 103 can be an EX_port while the peer port of FC router 128 can be an E port. As a result, topology information of network 100 is not propagated to FC fabric 150.

During operation, to communicate with target 144, server 114 creates an FCoE frame, and forwards the FCoE frame to ingress RBridge 103. RBridge 103 checks the encapsulated FC packet and identifies destination target 144 to be coupled to virtual RBridge 138. RBridge 103 recognizes the identifier of virtual RBridge to be a local identifier, extracts the FC packet from the FCoE frame, and forwards the FC packet to FC router 128. FC router 128, in turn, forwards the FC packet to FC router 124 coupled to target 144. Similarly, server 112 communicates with target 146 via RBridge 104, hybrid RBridge 101, and FC router 128.

In some embodiments, FC router 128 provides zoning support to targets 142, 144, and 146. Such zoning support is required to create logical partitions in FC fabric 150. A respective logical partition enables a selective communication between hybrid RBridges 101, 102, and 103, and targets 142, 144, and 146. In some embodiments, FC router 128 uses the logical partitions to present the targets as parts of different storage area networks. For example, if targets 142 and 144 belong to one zone while target 146 belongs to a different zone, FC router 128 presents a respective zone to hybrid RBridge 101, 102, and 103 as separate storage area networks. Consequently, hybrid RBridge 101, 102, and 103 create a virtual RBridge for targets 142 and 144, while another virtual RBridge for target 146.

Network Configuration

FIG. 2A illustrates an exemplary network configuration of a TRILL network coupled to virtualized FC routers, in accordance with an embodiment of the present invention. In this example, a TRILL network 200 includes TRILL RBridges 212, 214, and 216. Also included are RBridges 202 and 204, each with a number of Ethernet edge ports which can be coupled to end devices. For example, server 242 is coupled with RBridges 202 via an Ethernet edge port. Network 200 also includes hybrid RBridges 222 and 224, each with a number of edge ports for coupling end devices, and a number of FC ports for coupling FC routers. For example, RBridges 224 is coupled with server 244 via an Ethernet edge port and with FC router 232 via an FC EX_port, separating the topology information of network 200 from the FC domain. In some embodiments, hybrid RBridges 222 and 224 have only TRILL and FC ports, and not any Ethernet port. For example, hybrid RBridge 222 can operate without any Ethernet port because no end device is coupled with hybrid RBridge 222. RBridges in network 200 are interconnected with each other using TRILL ports.

Hybrid RBridge 222 creates a notification message presenting FC router 232 as virtual RBridge 234 and target 236 as a storage device coupled to virtual RBridge 234. Hybrid RBridge 222 then sends the notification message to all other RBridges in network. Upon receiving the message, other RBridges considers virtual RBridge 234 as a result RBridge coupled to hybrid RBridge 222. In some embodiments, network 200 is a fabric switch, wherein the notification message is created using an internal messaging service of the fabric switch.

Hybrid RBridges 222 and 224 in network 200 enables server 242 to communicate with target 236 in FC domain via an Ethernet port. During operation, to communicate with target 236, server 242 creates an FCoE frame, and forwards the FCoE frame to ingress RBridge 202 via an Ethernet interface. RBridge 202 checks the encapsulated FC packet and identifies destination target 236 to be coupled to virtual RBridge 234. RBridge 202 encapsulates the FCoE frame in a TRILL packet destined to virtual RBridge 234 and forwards the packet to one of the intermediate RBridges 212, 214, and 216. Because TRILL supports equal-cost multipath routing, a respective TRILL packet from RBridge 202 can be routed to hybrid RBridge 222 via any of RBridges 212, 214, and 216. Because the identifier of virtual RBridge 234 is a local identifier in hybrid RBridge, upon receiving the TRILL packet, hybrid RBridge 222 recognizes the packet to be destined to virtual RBridge 234. Hybrid RBridge 222 then extracts the FC packet and forwards the FC packet to FC router 232 coupled to target 236. Though other RBridges in network 200 perceives virtual RBridge 234 to be coupled to hybrid RBridge 222 via a TRILL port, hybrid RBridge 222 actually forwards the FC packet via a local FC EX_port.

In some embodiments, hybrid RBridges 222 and 224 can also forward non-FC traffic. Server 244 is coupled to hybrid RBridge 224 via an Ethernet port. To communicate with server 244, server 242 creates an Ethernet frame, and forwards the frame to ingress RBridge 202. RBridge 202 checks the destination of the frame and identifies destination server 244 to be coupled to hybrid RBridge 224. RBridge 202 encapsulates the frame in a TRILL packet destined to egress RBridge 224 and forwards the packet to RBridge 224 via one of the intermediate RBridges 212, 214, and 216. Upon receiving the TRILL packet, hybrid RBridge 224 recognizes the TRILL packet to be destined to itself, extracts the Ethernet frame, and forwards the Ethernet frame to server 244. In this way, a hybrid RBridge can forward Ethernet traffic as well as storage (e.g., FC) traffic.

FIG. 2B illustrates an exemplary network configuration of a TRILL network with hybrid RBridges coupled to servers and virtualized FC routers, in accordance with an embodiment of the present invention. In this example, a TRILL network 250 includes TRILL RBridges 262 and 264. Also included are hybrid RBridges 252 and 254, each with a number of edge ports for coupling end devices, and a number of FC ports for coupling FC routers. For example, RBridges 254 is coupled with server 282 via an Ethernet edge port and with FC router 272 via an FC EX_port, separating the topology information of network 250 from the FC domain. Hybrid RBridge 254 creates a notification message presenting FC router 272 as virtual RBridge 274 and target 276 as a storage device coupled to virtual RBridge 274. Hybrid RBridge 254 then sends the notification message to all other RBridges in network. In some embodiments, network 250 is a fabric switch, wherein the notification message is created using an internal messaging service of the fabric switch.

In some embodiments, hybrid RBridge 254 marks the MAC address of FC router 272 as a local MAC address. The MAC address of FC router 272 can be derived from the domain identifier of FC router 272. During operation, to communicate with target 276, server 282 creates an FCoE frame, and forwards the FCoE frame to hybrid RBridge 254 via an Ethernet port. Upon receiving the FCoE frame, hybrid RBridge 254 recognizes the MAC address of the FCoE frame to be a local MAC address and considers the FCoE frame to be destined to itself. As a result, hybrid RBridge 254 extracts the internal FC packet from the FCoE frame and promotes the frame to the upper layer. In the upper layer, hybrid RBridge 254 identifies target 276 as the destination of the FC packet and forwards the FC packet to FC router 272 coupled to target 276. In this way, when hybrid RBridge 254 is coupled to server 282, hybrid RBridge 254 can forward traffic to FC domain without encapsulating the FCoE frame in a TRILL packet. This reduces the load on inter-switch links of network 250.

However, hybrid RBridge 254 still presents FC router 272 as virtual RBridge 274 for other servers coupled to network 250. For example, to communicate with target 276, server 286 creates an FCoE frame and forwards the frame to RBridge 262. Upon receiving the FCoE frame, ingress RBridge 262 encapsulates the frame in a TRILL packet destined to virtual RBridge 274 and forwards the packet to hybrid RBridge 254. Hybrid RBridge 254 receives the TRILL packet, recognizes the TRILL packet to be destined to virtual RBridge 274, extracts the FC packet, and forwards the FC packet to FC router 272 coupled to target 276. Hence, FC router virtualization is required in network 250 even when a hybrid RBridge can forward traffic to FC domain without encapsulating an FCoE frame in a TRILL packet.

Priority-Based Flow Control

Priority-based flow control facilitates different classes of service for flow control of Ethernet frames. Such flow control enables control over a respective data flow on shared links. Priority-based flow control allows FCoE frames to receive a lossless service from a link shared with traditional lost-tolerant traffic. To facilitate priority-based flow control, whenever a hybrid RBridge learns about a target, the hybrid RBridge configures a class of service associated with the FCoE frames and propagates the configuration information to other RBridges in the TRILL network. FIG. 3A presents a flowchart illustrating the process of a hybrid RBridge distributing priority-based flow control configuration information to other RBridges in a TRILL network, in accordance with an embodiment of the present invention. The hybrid RBridge first receives class of service information associated with FC traffic from a locally coupled FC router (operation 302). The hybrid RBridge then creates a buffer dedicated for the FCoE frames belonging to the class of service (operation 304). In some embodiments, a network administrator can configure the priority-based flow control in the hybrid RBridge.

The hybrid RBridge then creates a notification message containing the priority-based flow control configuration for other RBridges in the TRILL network (e.g., other member switches in a fabric switch) (operation 306). This notification message can be an Ethernet control message. The hybrid RBridge encapsulates the notification message in a TRILL packet (operation 308) and sends the TRILL packet to other RBridges (operation 310). In some embodiments, the hybrid RBridge uses multicast to distribute the notification message.

FIG. 3B presents a flowchart illustrating the process of an RBridge self-configuring priority-based flow control based on received information from another RBridge in a TRILL network, in accordance with an embodiment of the present invention. Upon receiving a notification message from another RBridge (operation 352), the RBridge extracts the content from the notification message (operation 354). In some embodiments, the extraction of the content includes decapsulation of a TRILL header of an Ethernet frame. The RBridge then checks whether the notification message contains any priority-based flow control configuration (operation 356). In some embodiments, the RBridge examines a message type and an associated operation code of the message to determine whether the message pertains to priority-based flow control. If the message does not contain any priority-based flow control configuration, the RBridge takes action based on the content of the message (operation 358).

If the message contains priority-based flow control configuration, the RBridge identifies the class of service from the configuration (operation 362). In some embodiments, the RBridge identifies the class of service by identifying a priority value associated with the class in the message. The RBridge then dedicates a buffer for the class of service (operation 364) and initiates priority-based flow control for the class of service (operation 366). Such initiation can include, but is not limited to, monitoring the dedicated buffer, identifying a potential overflow, and calculating time to transmit frames, both current and incoming, from the buffer. In some embodiments, the RBridge uses priority-based flow control specified by Institute of Electrical and Electronics Engineers (IEEE) specification 802.1Qbb, “Priority-based Flow Control,” available at http://www.ieee802.org/1/pages/802.1bb.html, the disclosure of which is incorporated herein in its entirety.

In the example in FIG. 1, when hybrid RBridges 101, 102, and 103 learn about FC routers 122, 124, and 126, hybrid RBridges 101, 102, and 103 dynamically virtualize FC routers 122, 124, and 126, and present them as virtual RBridges 132, 134, and 136, respectively, by sending notification messages to RBridges 104, 105, 106, and 107. Upon receiving the notification messages, RBridges 104, 105, 106, and 107 dynamically learn about virtual RBridges 132, 134, and 136 as regular RBridges coupled to hybrid RBridges 101, 102, and 103, respectively. FIG. 4A presents a flowchart illustrating the process of a hybrid RBridge distributing virtualized FC connectivity information to other RBridges in a TRILL network, in accordance with an embodiment of the present invention. Upon detecting a new networking device (operation 402), the hybrid RBridge checks whether the detected device is an FC router (operation 404). If not, then the hybrid RBridge checks whether the detected device is an end device (operation 406). Note that the detected end device can be a server.

If the detected device is an end device, the hybrid RBridge creates a notification message identifying the end device (operation 424). In some embodiments, the hybrid RBridge detects the end device by learning the MAC address of the end device from a received frame and includes the learned MAC address in the notification message. If the detected device is not an FC router or an end device, the detected device is a new RBridge joining the TRILL network. The hybrid RBridge then creates a notification message identifying the RBridge (operation 422). In some embodiments, the hybrid RBridge detects the new RBridge from a received notification message from another RBridge.

If the identified networking device is an FC router (operation 404), the hybrid RBridge marks the layer-2 identifier of the FC router as a local identifier (operation 412). For example, if the identifier is a MAC address, the hybrid RBridge marks the MAC address of the FC router as a local MAC address. The hybrid RBridge then virtualizes the FC router as an RBridge (operation 414). In some embodiments, the RBridge identifier and the FC domain identifier have the same format, and the domain identifier of the FC router is used as the virtual RBridge identifier. Hybrid RBridge creates a notification message identifying the FC router as a (virtual) RBridge (operation 416). Note that the RBridges receiving the notification message recognize the FC router as a regular RBridge.

The hybrid RBridge can, optionally, learn about the target storage devices coupled to the FC router (operation 426) and associates the targets with the (virtual) RBridge in the notification message (operation 428). After creating a notification message identifying the detected networking device (operation 422, 424, or 428), the hybrid RBridge forwards the notification message to all other RBridges in the TRILL network. In some embodiments, the hybrid RBridge and the other RBridges in the TRILL network can be member switches of a fabric switch. The hybrid RBridge can use an internal messaging service of the fabric switch to create the notification message.

FIG. 4B presents a flowchart illustrating the process of an RBridge learning about other networking devices via a TRILL port, in accordance with an embodiment of the present invention. Upon receiving a notification message from a remote RBridge (operation 452), the RBridge checks whether the notification message contains information about one or more end devices (operation 454). If the notification contains end device information, the RBridge associates the end devices with the ingress RBridge of the message (operation 462). The RBridge then checks whether the ingress RBridge is a new RBridge (operation 464). A new RBridge is an RBridge for which the RBridge does not already have information stored in a local database. In some embodiments, the local database maintains the state of a fabric switch and is used to manage the member switches. If the notification message does not contain end device information (operation 454), the notification message contains information about a new RBridge. If the message contains information about a new RBridge (operation 454 or 464), the RBridge stores the new RBridges information in a local database (operation 456). Note that the new RBridge information can include the associated end device information as well.

Frame Processing

FIG. 5 presents a flowchart illustrating the process of an RBridge forwarding an Ethernet-encapsulated FC packet received from an end device, in accordance with an embodiment of the present invention. The RBridge first receives an FCoE frame from the end device via a local edge port (operation 502). The end device can be a server accessing data from a target storage device in a storage area network. The RBridge then examines the internal FC packet of the receiving FCoE frame to identify the target of the FC frame (operation 504) and identifies the RBridge coupled to the target (operation 506). Note that the identified RBridge can be a virtual RBridge. However, the RBridge considers the virtual RBridge as a regular RBridge and is not aware that the virtual RBridge is actually an FC router.

The RBridge encapsulates the received FCoE frame in a TRILL packet with an identifier of the virtual RBridge as the egress RBridge identifier (operation 508). The RBridge uses local forwarding information to identify the next hop RBridge and assigns the MAC address of the next hop RBridge as the destination address of the outer Ethernet header of the TRILL packet (operation 510). Note that the RBridge uses the same local forwarding information to determine routes for both FC and non-FC traffic. The RBridge forwards the TRILL packet to the next hop RBridge (operation 512). For the example in FIG. 2A, RBridge 202 encapsulates an FCoE frame for target 236 received from server 242 in a TRILL packet with an identifier of virtual RBridge 234 as the egress RBridge identifier. RBridge 202 identifies RBridge 212 as the next hop RBridge and assigns the MAC address of RBridge 212 as the destination address of the outer Ethernet header of the TRILL packet. RBridge 202 then forwards the TRILL packet to RBridge 212 based on the local forwarding information. Because the egress RBridge of the TRILL packet is RBridge 222, intermediate RBridge 212 does not examine the internal FC frame and forwards the TRILL packet to RBridge 222 using regular TRILL routing.

FIG. 6 presents a flowchart illustrating the process of a hybrid RBridge processing a TRILL packet, in accordance with an embodiment of the present invention. The hybrid RBridge first receives a TRILL packet via a TRILL port (operation 602). The TRILL port couples another RBridge via an inter-switch link. The hybrid RBridge checks whether the TRILL packet is for the local RBridge (i.e., the hybrid RBridge) (operation 604). If not, the TRILL packet is for another RBridge and the hybrid RBridge forwards the packet to the next hop RBridge (operation 628). If the TRILL packet is for the local RBridge, the hybrid RBridge extracts the internal frame from the TRILL packet (operation 606). The hybrid RBridge checks whether the extracted frame is for the local RBridge (operation 610). If the frame is for the local RBridge, the hybrid RBridge decapsulates the frame and the inner packet is promoted to the upper layer (operation 612).

In some embodiments, the hybrid RBridge marks the MAC address of the virtual RBridge (i.e., the FC router) as a local MAC address. Hence, the decapsulated packet can be an FC packet for an FC router coupled to the hybrid RBridge. The hybrid RBridge checks whether the packet is an FC packet (operation 614). If so, the hybrid RBridge identifies the target of the FC packet (operation 616) and the FC router coupled to the target (operation 618). The hybrid RBridge then forwards the FC packet to the FC router via a local FC port (operation 620). In some embodiments, the FC port is an EX_port. If the frame is not for the local RBridge (operation 610), the frame is for an end device coupled to the RBridge via a layer-2 interconnection. If the internal packet is not an FC packet (operation 614), the packet is for an end device coupled to the RBridge via a layer-3 interconnection. The RBridge identifies the end device (operation 622) and forwards the frame or packet to the end device (operation 624).

Exemplary RBride

FIG. 7 illustrates an exemplary RBridge, in accordance with an embodiment of the present invention. In this example, an RBridge 700 includes a number of communication ports 702, a flow control module 720, a device management module 730, an identification module 732, a packet processor 710, a forwarding module 740, and a storage 750. Forwarding module 720 makes local routing decision for a respective TRILL packet using a single routing protocol instance. The TRILL packet can contain an FCoE frame or a regular Ethernet frame (i.e., a non-FC frame). RBridge 700 can generate the TRILL or the FC packet using packet processor 710, or receive via one of the communication ports 702. Examples of the routing protocol can include, but are not limited to OSPF, IS-IS, and Distance-vector. Communication ports 702 can include one or more ports capable of receiving frames encapsulated in a TRILL header and/or an FC header. Packet processor 710 can process these frames.

In some embodiments, RBridge 700 may maintain a membership in a fabric switch, wherein RBridge 700 also includes a fabric switch management module 760. Fabric switch management module 760 maintains a configuration database in storage 750 that maintains the configuration state of a respective switch within the fabric switch. Fabric switch management module 760 maintains the state of the fabric switch, which is used to join other switches. Under such a scenario, communication ports 702 can include inter-switch communication channels for communication within a fabric switch. This inter-switch communication channel can be implemented via a regular communication port and based on any open or proprietary format.

During operation, RBridge 700 detects an FC router coupled to RBridge 700. RBridge 700 virtualizes the FC router and presents the FC router as a virtual RBridge to the rest of the TRILL network. Packet processor 710 constructs a notification message containing an RBridge identifier associated with the FC router. Device management module 730 operates in conjunction with packet processor 710 and checks whether a received TRILL packet has the RBridge identifier as the egress RBridge identifier. If so, device management module 730 terminates TRILL forwarding for the packet and packet processor 710 extracts an Ethernet frame from the TRILL packet. Device management module 730 identifies a MAC address of the FC router configured as a local address. If the MAC address of the Ethernet frame matches the MAC address of the FC router, packet processor 710 extracts an FC packet from the Ethernet frame.

Identification module 732 identifies a class of service associated with the FC router. Flow control module 720 then configures priority-based flow control associated with the class of service for RBridge 700. Packet processor 710 creates a notification message for a remote RBridge, which can be a member switch of a fabric switch, in response to identifying the class of service. The notification message can contain the class of service associated with the priority-based flow control configured for RBridge 700. When RBridge 700 learns information about a networking device, such as a MAC address, via one of the communication ports 702, packet processor 710 constructs a notification message for the remote RBridge containing the learned information.

Note that the above-mentioned modules can be implemented in hardware as well as in software. In one embodiment, these modules can be embodied in computer-executable instructions stored in a memory which is coupled to one or more processors in RBridge 700. When executed, these instructions cause the processor(s) to perform the aforementioned functions.

In summary, embodiments of the present invention provide a switch and a method for integrating a TRILL network with a storage area network. In one embodiment, the switch includes a packet processor and a device management module. During operation, the packet processor constructs a notification message containing a Transparent Interconnection of Lots of Links (TRILL) Routing Bridge (RBridge) identifier associated with a Fibre Channel router. The device management module operates in conjunction with the packet processor and terminates TRILL forwarding for a received TRILL packet with the RBridge identifier as an egress RBridge identifier.

The methods and processes described herein can be embodied as code and/or data, which can be stored in a computer-readable non-transitory storage medium. When a computer system reads and executes the code and/or data stored on the computer-readable non-transitory storage medium, the computer system performs the methods and processes embodied as data structures and code and stored within the medium.

The methods and processes described herein can be executed by and/or included in hardware modules or apparatus. These modules or apparatus may include, but are not limited to, an application-specific integrated circuit (ASIC) chip, a field-programmable gate array (FPGA), a dedicated or shared processor that executes a particular software module or a piece of code at a particular time, and/or other programmable-logic devices now known or later developed. When the hardware modules or apparatus are activated, they perform the methods and processes included within them.

The foregoing descriptions of embodiments of the present invention have been presented only for purposes of illustration and description. They are not intended to be exhaustive or to limit this disclosure. Accordingly, many modifications and variations will be apparent to practitioners skilled in the art. The scope of the present invention is defined by the appended claims.

Claims

1. A first switch, comprising: a plurality of ports comprising a first port capable of forwarding based on a first protocol in a network of interconnected switches and a second port capable of forwarding based on a second protocol;a packet processor configured to: construct a notification message indicating a physical switch identifier of a second switch, which operates based on the second protocol, as a second switch identifier of a third switch capable of operating based on the first protocol, wherein the physical switch identifier and the second switch identifier are of a same format;determine the first port as an output port for the notification message, wherein the first port corresponds to another switch; anddetermine the second port as an output port for a packet with the physical switch identifier as an egress switch identifier based on the physical switch identifier, wherein the second port corresponds to the second switch.

2. The first switch of claim 1, wherein the packet processor is further configured to extract the packet by decapsulating an encapsulation header associated with the first protocol, wherein the destination address of the encapsulation header corresponds to the third switch.

3. The first switch of claim 2, wherein the packet is a Fibre Channel packet, and wherein the second switch is a Fibre Channel router.

4. The first switch of claim 1, further comprising device management circuitry configured to: terminate forwarding for an Ethernet header of the packet, prior to determining the second port as an output port, by identifying a media access control (MAC) address of the second switch as a local address associated with the first switch.

5. The first switch of claim 1, further comprising: identification circuitry configured to identify a first class of service associated with the second switch based on the second protocol; andflow control circuitry configured to, corresponding to the first class of service, configure a second class of service based on priority-based flow control for the first switch.

6. The first switch of claim 5, wherein the packet processor is further configured to create a notification message for a remote switch in the network of interconnected switches in response to identifying the first class of service, wherein the notification message includes the second class of service configured for the first switch.

7. The first switch of claim 1, wherein the packet processor is further configured to construct a notification message for a remote switch in the network of interconnected switches, wherein the notification message includes a MAC address learned from a port in the plurality of ports.

8. The first switch of claim 1, further comprising forwarding circuitry configured to make a local routing decision for a second packet using a routing protocol instance based on the first protocol, wherein the second packet includes an inner packet destined to a remote switch outside of the network of interconnected switches.

9. The first switch of claim 8, wherein the routing protocol corresponds to one of: Open Shortest Path First (OSPF);Intermediate System to Intermediate System (IS-IS); andDistance-vector.

10. The first switch of claim 1, further comprising a fabric switch management module configured to maintain a membership in the network of interconnected switches, wherein the network of interconnected switches is identified by a fabric identifier.

11. A method, comprising: constructing, by a first switch, a notification message, wherein the first switch includes a plurality of ports comprising a first port capable of forwarding based on a first protocol in a network of interconnected switches and a second port capable of forwarding based on a second protocol;wherein the notification message indicates a physical switch identifier of a second switch, which operates based on the second protocol, as a second switch identifier of a third switch capable of operating based on the first protocol, wherein the physical switch identifier and the second switch identifier are of a same format; anddetermining the first port as an output port for the notification message, wherein the first port corresponds to another switch; anddetermining the second port as an output port for a packet with the physical switch identifier as an egress switch identifier based on the physical switch identifier, wherein the second port corresponds to the second switch.

12. The method of claim 11, wherein the method further comprises extracting the packet by decapsulating an encapsulation header associated with the first protocol, wherein the destination address of the encapsulation header corresponds to the third switch.

13. The method of claim 12, wherein the packet is a Fibre Channel packet, and wherein the second switch is a Fibre Channel router.

14. The method of claim 11, further comprising: terminating forwarding for an Ethernet header of the packet, prior to determining the second port as an output port, by identifying a media access control (MAC) address of the second switch as a local address associated with the first switch.

15. The method of claim 11, further comprising: identifying a first class of service associated with the second switch based on the second protocol; andcorresponding to the first class of service, configuring a second class of service based on priority-based flow control for the first switch.

16. The method of claim 15, further comprising creating a notification message for a remote switch in the network of interconnected switches in response to identifying the first class of service, wherein the notification message includes the second class of service configured for the first switch.

17. The method of claim 11, further comprising constructing a notification message for a remote switch in the network of interconnected switches, wherein the notification message includes a MAC address learned from a port in the plurality of ports of the first switch.

18. The method of claim 11, further comprising making a local routing decision for a second packet using a routing protocol instance based on the first protocol, wherein the second packet includes an inner packet destined to a remote switch outside of the network of interconnected switches.

19. The method of claim 18, wherein the routing protocol corresponds to one of: Open Shortest Path First (OSPF);Intermediate System to Intermediate System (IS-IS); andDistance-vector.

20. The method of claim 11, further comprising maintaining a membership in the network of interconnected switches, wherein the network of interconnected switches is identified by a fabric identifier.