Intelligent network interfaced device and system for accelerated communication

• US 6,427,173 B1
• Filed: 12/15/1999
• Issued: 07/30/2002
• Est. Priority Date: 10/14/1997
• Status: Expired due to Term

• Current Assignee: Alacritech, Inc.
• Original Assignee: Alacritech, Inc.
• Inventors: Higgen, David A., Philbrick, Clive M., Boucher, Laurence B., Craft, Peter K., Blightman, Stephen E. J., Starr, Daryl D.
• Primary Examiner(s): Maung, Zarni
• Application Number:
  US09/464,283
• Time in Patent Office: 958 Days
• Field of Search: 709/230, 709/238, 709/236, 709/228, 709/232, 709/250, 709/225
• US Class Current: 709/238
• CPC Class Codes: G06F 5/10 having a sequence of storage locations each being individually accessible for both enqueue and dequeue operations, e.g. using random access memory G06F5/065 takes precedence
• CPC Class Codes: H04L 49/90 Buffering arrangements
• CPC Class Codes: H04L 49/9063 Intermediate storage in different physical parts of a node or terminal
• CPC Class Codes: H04L 67/62 Establishing a time schedule for servicing the requests
• CPC Class Codes: H04L 67/63 Routing a service request depending on the request content or context
• CPC Class Codes: H04L 69/08 Protocols for interworking; Protocol conversion
• CPC Class Codes: H04L 69/12 Protocol engines
• CPC Class Codes: H04L 69/16 Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
• CPC Class Codes: H04L 69/161 Implementation details of TCP/IP or UDP/IP stack architecture; Specification of modified or new header fields
• CPC Class Codes: H04L 69/162 involving adaptations of sockets based mechanisms secure socket layer H04L63/168
• CPC Class Codes: H04L 69/163 In-band adaptation of TCP data exchange; In-band control procedures
• CPC Class Codes: H04L 69/166 IP fragmentation; TCP segmentation
• CPC Class Codes: H04L 69/22 Parsing or analysis of headers
• CPC Class Codes: H04L 69/32 Architecture of open systems interconnection [OSI] 7-layer type protocol stacks, e.g. the interfaces between the data link level and the physical level
• CPC Class Codes: H04L 69/325 in the network layer [OSI layer 3], e.g. X.25 H04L69/16 takes precedence
• CPC Class Codes: H04L 69/326 in the transport layer [OSI layer 4] H04L69/16 takes precedence
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

An intelligent network interface card (INIC) or communication processing device (CPD) works with a host computer for data communication. The device provides a fast-path that avoids protocol processing for most messages, greatly accelerating data transfer and offloading time-intensive processing tasks from the host CPU. The host retains a fallback processing capability for messages that do not fit fast-path criteria, with the device providing assistance such as validation even for slow-path messages, and messages being selected for either fast-path or slow-path processing. A context for a connection is defined that allows the device to move data, free of headers, directly to or from a destination or source in the host. The context can be passed back to the host for message processing by the host. The device contains specialized hardware circuits that are much faster at their specific tasks than a general purpose CPU. A preferred embodiment includes a trio of pipelined processors devoted to transmit, receive and utility processing, providing full duplex communication for four Fast Ethernet nodes. The device provides a fast-path that avoids protocol processing for most messages, greatly accelerating data transfer and offloading time-intensive processing tasks from the host CPU. The host retains a fallback processing capability for messages that do not fit fast-path criteria, with the device providing assistance such as validation even for slow-path messages, and messages being selected for either fast-path or slow-path processing. A context for a connection is defined that allows the device to move data, free of headers, directly to or from a destination or source in the host. The context can be passed back to the host for message processing by the host.

Claims

1. A device for communication between a network and a host computer, the device being coupled to the network and being coupled to the host computer, the host computer having a protocol stack, the device comprising:

a receive sequencer that analyzes headers of packets received onto the device from the network, the receive sequencer generating for each of the packets an associated summary, each of the summaries being indicative of whether the associated packet is a fast-path candidate;

a memory that stores the summaries; and

a processor that obtains the summaries from the memory and determines, for each summary indicating that the associated packet is a fast-path candidate, whether the associated packet meets a plurality of fast-path criteria, and wherein if the processor determines that a packet meets the plurality of fast-path criteria then the processor performs transport and network layer processing of the packet on the device such that the protocol stack of the host computer does not perform the transport or network layer processing of the packet, whereas if the processor determines that the packet does not meet the plurality of fast-path criteria then the processor does not perform the transport and network layer processing such that the transport and network layer processing is done by the protocol stack of the host computer.

2. The device of claim 1, wherein only packets utilizing a particular transport layer protocol and a particular network layer protocol can be fast-path candidates, and wherein each packet received onto the host utilizes a transport layer protocol and a network layer protocol, the summary of a packet being indicative of whether the transport layer protocol utilized by the packet is the particular transport layer protocol, the summary of the packet also being indicative of whether the network layer protocol utilized by the packet is the particular network layer protocol.

3. The device of claim 1, wherein a plurality of connection contexts is maintained on the device, and wherein the determination by the processor that the packet meets the plurality of fast-path criteria involves determining that a connection context associated with the packet is one of the plurality of connection contexts maintained on the device.

4. The device of claim 1, wherein a message comprising a plurality of packets is received onto the device, each of the packets of the message having a header portion and a data portion, the receive sequencer generating for each of the packets of the message a summary indicating that the packet is a fast-path candidate, the processor determining that each of the packets of the message meets the plurality of fast-path criteria, the data portions of the packets of the message being transferred to the host computer and stored as a block of data.

5. The device of claim 4, wherein the block of data contains no transport layer headers and contains no network layer headers.

6. The device of claim 4, wherein the processor does not attempt to determine whether a particular packet meets the plurality of fast-path criteria if the summary for the particular packet indicates that the particular packet is not a fast-path candidate.

7. A device for communication between a network and a host computer, the device being coupled to the network and being coupled to the host computer, the host computer having a protocol stack, the device comprising:

a receive sequencer that analyzes headers of packets received onto the device from the network, the receive sequencer generating for each of the packets an associated summary;

a memory that stores the summaries; and

a mechanism that obtains the summaries from the memory and determines, for each summary, whether;

     1) the packet associated with the summary is processed by the protocol stack of the host, or

     2) data contained in the packet is delivered to a destination in the host without protocol processing by the host.

8. The device of claim 7, wherein the summary contains an attention bit indicative of whether the associated packet is a candidate for being delivered to the destination in the host without protocol processing by the host.

9. The device of claim 8, wherein the associated packet has a transport layer protocol and a network layer protocol, the packet not being a candidate for being delivered to the destination in the host without protocol processing by the host:

1) if the transport layer protocol of the packet is not a predetermined transport layer protocol, and

     2) if the network layer protocol of the packet is not a predetermined network layer protocol.

10. The device of claim 7, wherein the mechanism contains a processor that checks the packet for an exception condition prior to the packet being delivered to the destination in the host without protocol processing by the host.

11. The device of claim 7, wherein a session layer message comprises a first packet and a second packet, each of the first and second packets comprising a header and data, the destination in the host being a portion of memory, and wherein the mechanism delivers the data of the first packet and the data of the second packet to the destination such that the data from the first packet and the data of the second packet are stored in the destination without any header being stored between the data of the first packet and the data of the second packet.

12. The device of claim 11, wherein the first packet is a transport layer packet and wherein the second packet is a transport layer packet, there being no transport layer header stored in the destination between the data of the first packet and the data of the second packet, there being no network layer header stored in the destination between the data of the first packet and the data of the second packet.

13. The device of claim 7, wherein the packet associated with the summary has an associated context, the mechanism determining whether the associated context is one of a plurality of contexts, wherein if the associated context is not one of the plurality of contexts then the packet associated with the summary is processed by the protocol stack of the host.

14. The device of claim 7, wherein each packet received onto the device has one and only one associated context, and wherein the destination is associated with one and only one context, wherein only data of packets having a particular context are stored in the destination associated with that particular context.

15. A system for communicating a TCP data packet through a network from a first computer to a second computer, the TCP data packet being communicated from the first computer to the second computer via a TCP/IP connection, the system comprising:

a device coupled to the network, the device comprising a first portion and a second portion, the first portion performing initial TCP processing on the TCP data packet as the TCP data packet is received onto the device from the network, the second portion being a processor that performs subsequent TCP processing on the TCP data packet; and

a protocol processing stack disposed in the second computer and configured for passing control of the TCP/IP connection to the device, the control being passed from the protocol processing stack to the device, the TCP data packet being generally processed by the device instead of the protocol processing stack when the device is controlling the TCP/IP connection.

16. The system of claim 15, wherein the device is a communication processing device (CPD) integrated into the second computer.

17. The system of claim 15, wherein the device is an intelligent network interface card (INIC) that is added to the second computer.

18. The system of claim 15, wherein the first portion performs TCP checksum validation.

19. The system of claim 15, wherein the first portion sets a bit to indicate that the TCP data packet is a fast-path candidate, the bit being read by the second portion.

20. The system of claim 15, wherein the device performs substantially all TCP protocol receive processing on the TCP data packet and the protocol processing stack performs substantially no TCP protocol receive processing on the TCP data packet.