Method to implement an L4-L7 switch using split connections and an offloading NIC
View Patent ↗A method of operating intelligent network interface circuitry includes the network interface circuitry coupling a core processor to a network to facilitate communication over the network between the core processor and at least a first peer and a second peer. A first connection connects to the first peer and a second connection connects to the second peer. The network interface circuitry receives data packets from the first peer via the network on the first connection, according to a first particular protocol. The network interface circuitry processes the received data, including associating, with the second connection, data that is at least a portion of the data packets received on the first connection, such that the data received by the intelligent network interface circuitry on the first connection is switched to be outgoing from the intelligent network interface circuitry on the second connection, according to a second particular protocol.
1. A method of operating intelligent network interface circuitry, wherein the network interface circuitry couples a core processor to a network to facilitate communication over the network between the core processor and at least a first peer and a second peer, wherein a first connection connects to the first peer and a second connection connects to the second peer, a plurality of connections, including the second connection, connecting to a plurality of separate peers, the method comprising:
by the network interface circuitry,
receiving data packets from the first peer via the network on the first connection, according to a first particular protocol; and
processing the received data, including associating, with the second connection, data that is at least a portion of the data packets received on the first connection, such that the data received by the intelligent network interface circuitry on the first connection is switched to be outgoing from the intelligent network interface circuitry on the second connection, according to a second particular protocol
wherein
the first particular protocol operates at no higher than layer 4, and
processing the received data includes processing the packets at higher than layer 4, wherein the associating is based on a result of processing the layers higher than layer 4 to determine which of the plurality of connections is the second connection.
2. The method of claim 1 , wherein:
the network interface circuitry includes a pipelined processor circuitry configured to process data received by the network interface circuitry from the core for transmission to a peer via one of the connections and also to process data received by the network interface circuitry via one of the connections for receipt by the core; and
associating, with the second connection, data that is at least a portion of the data packets received on the first connection includes generating data by the core processor, to be delivered on the second connection.
3. The method of claim 1 , wherein:
associating, with the second connection, data that is at least a portion of the data packets received on the first connection includes placing the data in a transmit buffer associated with the second connection.
4. The method of claim 3 , wherein:
the data is placed in the transmit buffer associated with the second connection without storing the data in a receive buffer associated with the first connection.
5. The method of claim 3 , wherein:
placing the data in a transmit buffer associated with the second connection includes appending the data to data already in the transmit buffer associated with the second connection.
6. The method of claim 1 , wherein:
the data received from the first peer via the network on the first connection is payload data included in data packets received from the first peer associated with the first connection;
the first connection and the second connection are full duplex, and
the method further comprises, by the intelligent network interface circuitry
receiving data packets from the second peer on the second connection;
associating, with the first connection, the data packets received on the second connection, such that the data incoming to the intelligent network interface circuitry on the second connection is switched to be outgoing from the intelligent network interface circuitry on the first connection.
7. The method of claim 1 , wherein:
processing at least a portion of data packets received on the first connection to determine which of the plurality of connections is the second connection includes
the network interface circuitry providing the at least a portion of the data packet, corresponding to higher than layer 4 and received on the first connection to the core processor; and
the core processor provides an indication to the network interface circuitry of the determination of which of the plurality of connections is the second connection.
8. The method of claim 7 , wherein:
the network interface circuitry providing the at least a portion of the data packet received on the first connection to the core processor includes delineating at least one protocol data unit in the data received on the first connection; and
the at least a portion of the data packet received on the first connection is at least a portion of the at least one protocol data unit delineated by the network interface circuitry.
9. The method of claim 1 , wherein:
the network interface circuitry determines which of the plurality of connections is the second connection by processing the at least a portion of the data packet, corresponding to higher than layer 4 and received on the first connection in view of control information associated with the connections.
10. The method of claim 1 , wherein:
the portion of the data packets processed for associating the received data with the second connection includes a portion of the data packets associated with layer 5 to 7 packet headers.
11. The method of claim 10 , wherein:
the portion of the data packets associated with layer 5 to 7 packet headers includes iSCSI packet headers.
12. The method of claim 11 , wherein:
the plurality of separate peers to which the plurality of connections connect are storage controllers.
13. The method of claim 1 , wherein:
the processing of at least a portion of data packets received on the first connection to determine which of the plurality of connections is the second connection is on a per Protocol Data Unit basis in at least one of layers 5-7.
14. The method of claim 13 , wherein:
the portion of the data packets associated with layer 5 to 7 packet headers includes iSCSI packet headers.
15. The method of claim 1 , further comprising:
performing supplemental processing on the data provided on the second connection, not associated with switching the data.
16. The method of claim 15 , wherein the supplemental processing includes verifying the integrity of the received data.
17. The method of claim 16 , including:
inserting, into the received data, a result of computing the integrity of the received data.
18. The method of claim 15 , wherein the supplemental processing includes at least one of encryption/decryption and compression/decompression.
19. The method of claim 1 , wherein:
the first protocol and the second protocol are variants of the same protocol with environment specific optimizations.
20. The method of claim 1 , wherein:
the first protocol and the second protocol are variants of the same protocol configured differently.
21. A method of operating intelligent network interface circuitry, wherein the network interface circuitry couples a core processor to a network to facilitate communication over the network between the core processor and at least a first peer and a second peer, wherein a first connection connects to the first peer and a second connection connects to the second peer, the method comprising:
by the network interface circuitry,
receiving data packets from the first peer via the network on the first connection, according to a first particular protocol; and
processing the received data, including associating, with the second connection, data that is at least a portion of the data packets received on the first connection, such that the data received by the intelligent network interface circuitry on the first connection is switched to be outgoing from the intelligent network interface circuitry on the second connection, according to a second particular protocol, wherein:
the network interface circuitry includes a pipelined processor circuitry configured to process data received by the network interface circuitry from the core for transmission to a peer via one of the connections and also to process data received by the network interface circuitry via one of the connections for receipt by the core; and
associating, with the second connection, data that is at least a portion of the data packets received on the first connection includes generating data by the core processor, to be delivered on the second connection; and
wherein the pipelined processor circuitry includes
ingress form packet circuitry configured to form ingress packets, from data received from the network, to provide to the core;
egress form packet circuitry configured to form egress packets, from data received from the core, to provide to the network;
intercept and redirect circuitry to selectively intercept packets from the ingress form packet circuitry, formed by the ingress form packet circuitry based on data received on the first connection, and to provide a message to the pipelined processor circuitry simulating a message from the core instructing the pipelined processor circuitry to provide the data of the ingress packets, from the ingress form packet circuitry based on data received on the first connection, to be redirected to the egress form packet circuitry to form egress packets to be delivered on the network on the second connection.
22. The method of claim 21 , wherein:
an identification of the second connection is stored in the network interface circuitry in a control block associated with the first connection.
23. The method of claim 21 , wherein:
the first particular protocol operates at no higher than layer 4, and
processing the received data includes processing the packets at higher than layer 4, wherein the associating is based on a result of processing the layers higher than layer 4.