PACKET SWITCHING BASED ON GLOBAL IDENTIFIER

• US 20120314706A1
• Filed: 06/07/2011
• Published: 12/13/2012
• Est. Priority Date: 06/07/2011
• Status: Active Grant

• Current Assignee: Mellanox Technologies TLV Ltd. (NVIDIA Corporation)
• Original Assignee: Mellanox Technologies Limited (NVIDIA Corporation)
• Inventors: Kagan, Michael, Crupnicoff, Diego, Liss, Liran
• Granted Patent: US 8,842,671 B2
• Time in Patent Office: Days
• US Class Current: 370/392
• CPC Class Codes: H04L 49/358 Infiniband Switches
• CPC Class Codes: H04L 49/602 Multilayer or multiprotocol switching, e.g. IP switching

Abstract

A communication method in a network operating in accordance with a standard that allocates a given number of bits m for layer-2 addressing of nodes in the network. The method includes accepting at a layer-2 switch in the network an assignment to one or more nodes in the network of respective layer-2 extended addresses, each including n=m+k bits, k>0. A given data packet is received at the switch for forwarding. The given data packet includes a layer-2 destination address and a layer-3 destination address in accordance with the standard. The layer-3 destination address includes t bits, t≧k. The given data packet is forwarded from the switch to one of the nodes by reading from the given data packet and combining the layer-2 destination address and k bits from the layer-3 destination address so as to reconstruct the n bits of the extended layer-2 address of the one of the nodes.

Claims

1. A method for communication in a network operating in accordance with a standard that allocates a given number of bits m for layer-2 addressing of nodes in the network, the method comprising:

accepting at a layer-2 switch in the network an assignment to one or more nodes in the network of respective layer-2 extended addresses, each comprising n=m+k bits, k>

0;

receiving at the switch a given data packet for forwarding, the given data packet comprising a layer-2 destination address and a layer-3 destination address in accordance with the standard, the layer-3 destination address comprising t bits, t≧

k; and

forwarding the given data packet from the switch to one of the nodes by reading from the given data packet and combining the layer-2 destination address and k bits from the layer-3 destination address so as to reconstruct the n bits of the extended layer-2 address of the one of the nodes.

2. The method according to claim 1, wherein the standard comprises an Infiniband standard, wherein the layer-2 destination address comprises a Destination Local Identifier (DLID), and wherein the layer-3 destination address comprises a Destination Global Identifier (DGID).

3. The method according to claim 1, and comprising receiving at the switch an additional data packet, which is to be forwarded to an additional node in the same subnet as the one of the nodes, wherein the additional data packet comprises the same layer-2 destination address as the given data packet but differs from the given data packet in at least one of the k bits of the layer-3 destination address, and forwarding the additional data packet based on the layer-2 destination address and the k bits of the layer-3 destination address of the additional data packet.

4. The method according to claim 1, and comprising defining in the switch, individually for each layer-2 destination address, whether the layer-2 destination address supports layer-2 forwarding using the n bits of the extended layer-2 address.

5. The method according to claim 4, and comprising forwarding an input data packet by:

when the input data packet supports forwarding using the n bits, forwarding the input data packet based on the n bits of the extended layer-2 address; and

when the input data packet does not support forwarding using the n bits, forwarding the input data packet based on the m bits of the layer-2 destination address.

6. The method according to claim 1, wherein the layer-3 destination address comprises a subnet prefix that identifies a subnet to which the one of the nodes belongs, and an interface identifier that identifies the one of the nodes within the subnet, and wherein the k bits are read from the interface identifier.

7. The method according to claim 1, wherein the one of the nodes comprises a virtual node.

8. The method according to claim 1, wherein the layer-2 switch comprises a virtual switch.

9. The method according to claim 1, wherein the layer-2 destination address and the layer-3 destination address of the given data packet comprise multicast addresses, such that the given data packet is addressed to multiple nodes in the network.

10. A layer-2 network switch, comprising:

one or more ports for communicating with a network operating in accordance with a standard that allocates a given number of bits m for layer-2 addressing of nodes in the network; and

circuitry, which is configured to accept an assignment to one or more nodes in the network of respective extended layer-2 addresses, each comprising n=m+k bits, k>

0, to receive via one of the ports a given data packet for forwarding, the given data packet comprising a layer-2 destination address and a layer-3 destination address in accordance with the standard, the layer-3 destination address comprising t bits, t≧

k, and to forward the given data packet from the switch to one of the nodes by reading from the given data packet and combining the layer-2 destination address and k bits from the layer-3 destination address so as to reconstruct the n bits of the extended layer-2 address of the one of the nodes.

11. The switch according to claim 10, wherein the standard comprises an Infiniband standard, wherein the layer-2 destination address comprises a Destination Local Identifier (DLID), and wherein the layer-3 destination address comprises a Destination Global Identifier (DGID).

12. The switch according to claim 10, wherein the circuitry is configured to receive an additional data packet, which is to be forwarded to an additional node in the same subnet as the one of the nodes, wherein the additional data packet comprises the same layer-2 destination address as the given data packet but differs from the given data packet in at least one of the k bits of the layer-3 destination address, and to forward the additional data packet based on the layer-2 destination address and the k bits of the layer-3 destination address of the additional data packet.

13. The switch according to claim 10, wherein the circuitry is configured to hold a definition, individually for each layer-2 destination address, whether the layer-2 destination address supports layer-2 forwarding using the n bits of the extended layer-2 address.

14. The switch according to claim 13, wherein the circuitry is configured to forward an input data packet by:

when the input data packet supports forwarding using the n bits, forwarding the input data packet based on the n bits of the extended layer-2 address; and

when the input data packet does not support forwarding using the n bits, forwarding the input data packet based on the m bits of the layer-2 destination address.

15. The switch according to claim 10, wherein the layer-destination address comprises a subnet prefix that identifies a subnet to which the one of the nodes belongs, and an interface identifier that identifies the one of the nodes within the subnet, and wherein the circuitry is configured to read the k bits from the interface identifier.

16. The switch according to claim 10, wherein the one of the nodes comprises a virtual node.

17. The switch according to claim 10, wherein the ports comprise virtual ports.

18. The switch according to claim 10, wherein the layer-2 destination address and the layer-3 destination address of the given data packet comprise multicast addresses, such that the given data packet is addressed to multiple nodes in the network.

19. A network interface, comprising:

one or more ports for communicating with a network operating in accordance with a standard that allocates a given number of bits m for layer-2 addressing of nodes in the network; and

a virtual layer-2 network switch configured to accept an assignment to one or more nodes in the network of respective extended layer-2 addresses, each comprising n=m+k bits, k>

0, to receive via one of the ports a given data packet for forwarding, the given data packet comprising a layer-2 destination address and a layer-3 destination address in accordance with the standard, the layer-3 destination address comprising t bits, t≧

k, and to forward the given data packet from the switch to one of the nodes by reading from the given data packet and combining the layer-2 destination address and k bits from the layer-3 destination address so as to reconstruct the n bits of the extended layer-2 address of the one of the nodes.

20. The network interface according to claim 19, wherein the standard comprises an Infiniband standard, wherein the layer-2 destination address comprises a Destination Local Identifier (DLID), and wherein the layer-3 destination address comprises a Destination Global Identifier (DGID).

21. The network interface according to claim 19, wherein the virtual switch holds a definition, individually for each layer-2 destination address, whether the layer-2 destination address supports layer-2 forwarding using the n bits of the extended layer-2 address.