I/O protocol for highly configurable multi-node processing system

• US 6,105,122 A
• Filed: 02/06/1998
• Issued: 08/15/2000
• Est. Priority Date: 02/06/1998
• Status: Expired due to Term

• Current Assignee: Teradata US, Inc. (Teradata Corporation)
• Original Assignee: NCR Corporation (NCR Voyix Corporation)
• Inventors: Muller, P. Keith, Chow, Kit M.
• Primary Examiner(s): Asta, Frank J.
• Assistant Examiner(s): Prieto, Beatriz
• Application Number:
  US09/020,199
• Time in Patent Office: 921 Days
• Field of Search: 395/500, 395/200, 395/250, 395/600, 395/700, 395/800, 395/172, 395/152, 395/500.47, 364/228.9, 364/228.4, 709/221, 709/224, 709/239, 709/217, 709/215, 710/10, 710/36, 710/131, 714/1-3, 714/25, 714/30, 714/31, 370/217, 370/244, 370/351, 712/43, 711/147, 340/825.04
• US Class Current: 712/1
• CPC Class Codes: G06F 15/17 using an input/output type connection, e.g. channel, I/O port
• CPC Class Codes: G06F 15/17368 non hierarchical topologies
• CPC Class Codes: G06F 21/6218 to a system of files or objects, e.g. local or distributed file system or database
• CPC Class Codes: H04L 12/56 Packet switching systems
• CPC Class Codes: H04L 63/0823 using certificates cryptographic mechanisms or cryptographic arrangements for entity authentication involving certificates H04L9/3263

Abstract

A method for transferring data from a first node to a second node in a multi-processor system is described. The multi-processor system comprises a plurality of nodes coupled to an interconnect fabric via an interconnect fabric interface, the nodes comprising a compute node and an I/O node, the I/O node coupled to a plurality of data storage devices. The method comprises the steps of generating a I/O request packet in a first node in response to an I/O request from an application executing in a first node, transmitting the data request packet to the second node via the interconnect fabric, and executing the destination interconnect channel program to extract the debit ID to transfer the data request to the second node buffer. In one embodiment, the I/O packet includes a data transfer request, an interconnect destination channel program, a first debit ID designating a second node buffer where the data request will be transmitted, and a first credit ID designating a first node buffer where the data responsive to the I/O request will be transmitted.

Claims

1. A method of transferring data from a first node to a second node in a multi-processor system comprising a plurality of nodes coupled to an interconnect fabric, the nodes comprising a plurality of compute nodes and a plurality of I/O nodes, each I/O node coupled to a plurality of data storage devices, the method comprising the steps of:

generating a I/O request packet in a first node in response to an I/O request from an application executing in the first node, the I/O packet having

a data transfer request;

an interconnect destination-side channel program;

a first debit ID designating a second node buffer where the data request will be transmitted; and

a first credit ID designating a first node buffer where the data responsive to the I/O request will be transmitted;

transmitting the data request packet to a second node via the interconnect fabric; and

executing the destination-side interconnect channel program to extract the first debit ID to transfer the data request to the second node buffer.

2. The method of claim 1, further comprising the steps of:

receiving a message signaling completion of the I/O request from the data storage device in the second node;

generating an I/O completion packet, the I/O completion packet comprising

a second debit ID corresponding to the first credit ID;

data responsive to the data transfer request; and

transmitting the I/O completion packet to the first node via the interconnect fabric.

3. The method of claim 2, wherein the I/O completion packet comprises a replacement credit ID designating a second buffer in the second node if the second node buffer is not free.

4. The method of claim 2, wherein the method further comprises the step of:

executing a send-side interconnect channel program to extract the second debit ID;

validating the second debit ID;

transferring the data responsive to the data transfer request to the first node buffer when the second debit ID is validated.

7. The apparatus of claim 2, wherein the I/O completion packet comprises a replacement credit ID designating a second buffer in the second node if the second node buffer is not free.

5. An apparatus for transferring data from a first node in a multi-processor system comprising a plurality of nodes coupled to an interconnect fabric, the nodes comprising a plurality of compute nodes and a plurality of I/O nodes, each I/O node coupled to a plurality of data storage devices, the apparatus comprising:

means for generating an I/O request packet in a first node in response to an I/O request from an application executing in the first node, the I/O packet having

a data transfer request;

an interconnect destination channel program;

a first debit ID designating a second node buffer where the data request will be transmitted;

a first credit ID designating a first node buffer where data responsive to the I/O request will be transmitted;

means for transmitting the data request packet to a second node via the interconnect fabric; and

means for executing the destination interconnect channel program to extract the debit ID to transfer the data request to the second node buffer.

6. The apparatus of claim 5, further comprising:

means for receiving a message signaling completion of the I/O request from the data storage devices in the second node;

means for generating an I/O completion packet, the I/O completion packet comprising a second debit ID corresponding to the first credit ID and data responsive to the data transfer request; and

means for transmitting the I/O completion packet to the first node via the interconnect fabric.

8. The apparatus of claim 6, wherein the apparatus further comprises:

means for executing a send-side interconnect channel program to extract the second debit ID;

means for validating the second debit ID; and

means for transferring the data responsive to the data transfer request to the first node buffer when the second debit ID is validated.

9. A program storage device, readable by a computer, tangibly embodying one or more programs of instructions executable by the computer to perform method steps of transferring data from a first node to a second node in a multi-processor system comprising a plurality of nodes coupled to an interconnect fabric, the nodes comprising a plurality of compute nodes and a plurality of I/O nodes, each I/O node coupled to a plurality of data storage devices, the method comprising the steps of:

generating a I/O request packet in a first node in response to an I/O request from an application executing in the first node, the I/O packet having

a data transfer request;

an interconnect destination-side channel program;

a first debit ID designating a second node buffer where the data request will be transmitted; and

a first credit ID designating a first node buffer where the data responsive to the I/O request will be transmitted;

transmitting the data request packet to a second node via the interconnect fabric; and

executing the destination-side interconnect channel program to extract the debit ID to transfer the data request to the second node buffer.

10. The program storage device of claim 9, wherein the method steps further comprise the steps of:

receiving a message signaling completion of the I/O request from the data storage devices in the second node;

generating an I/O completion packet, the I/O completion packet comprising a second debit ID corresponding to the first credit ID and data responsive to the data transfer request; and

transmitting the I/O completion packet to the first node via the interconnect fabric.

11. The program storage device of claim 10, wherein the I/O completion packet comprises a replacement credit ID designating a second buffer in the second node if the second node buffer is not free.

12. The program storage device of claim 10, wherein the method steps further comprise the method steps of:

executing a send-side interconnect channel program to extract the second debit ID;

validating the second debit ID; and

transferring the data responsive to the data transfer request to the first node buffer when the second debit ID is validated.

13. A data structure for transceiving payload data from a first node to a second node in a multiprocessor system, comprising:

a payload portion, having payload data;

a header portion, having

an interconnect channel program for processing the data packet at a second node interface having a second node buffer;

a debit field for storing a first ID designating a location in the second node buffer where an I/O request will be transmitted; and

a credit field for storing a second ID designating a location in a first node buffer where data responsive to the I/O request will be transmitted.

14. The data structure of claim 13, further comprising a second interconnect channel program for processing an I/O completion packet in a first node interface coupled to the first node buffer.

15. The data structure of claim 13, wherein the first ID grants the first node rights to use the second node buffer.

16. The data structure of claim 13, wherein the second ID grants the second node rights to use the first node buffer.

17. A method of transferring data from a first node to a second node in a multi-processor system comprising a plurality of nodes coupled to an interconnect fabric, the nodes comprising a plurality of compute nodes and a plurality of I/O nodes, each I/O node coupled to a plurality of data storage devices, the method comprising the steps of:

generating a I/O request packet in a first node in response to an I/O request from an application executing in the first node, the I/O packet having

a data transfer request;

an interconnect destination-side channel program;

a first debit ID designating a second node buffer where the data request will be transmitted; and

a first credit ID designating a first node buffer where the data responsive to the I/O request will be transmitted; and

transmitting the data request packet to a second node via the interconnect fabric.

18. The method of claim 17, wherein the interconnect destination channel program comprises instructions for executing the destination-side interconnect channel program to extract the debit ID to transfer the data request to the second node buffer.