System and method for flow control in a network

• US 20060031519A1
• Filed: 04/30/2004
• Published: 02/09/2006
• Est. Priority Date: 04/30/2004
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Gardner, Edward A., Helliwell, Richard Perham, Wrenn, Richard Fitzhugh
• Granted Patent: US 7,627,627 B2
• Time in Patent Office: Days
• US Class Current: 709/227
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/19 at layers above the network layer network arrangements for networked applications for scheduling or organising the servicing of application requests H04L67/60
• CPC Class Codes: H04L 47/27 Evaluation or update of window size, e.g. using information derived from acknowledged [ACK] packets
• CPC Class Codes: H04L 67/133 Protocols for remote procedure calls [RPC]
• CPC Class Codes: H04L 67/2876 Pairs of inter-processing entities at each side of the network, e.g. split proxies
• CPC Class Codes: H04L 67/62 Establishing a time schedule for servicing the requests
• CPC Class Codes: H04L 69/28 Timers or timing mechanisms used in protocols

Abstract

A method for controlling command message flow in a network including a server and a client. A command window, comprising a maximum number of command messages that may be outstanding at the server, is included in messages sent from the server to the client. The value of the command window at the server is modified in accordance with available server resources for receiving command messages. When there are insufficient resources at the server to process one of the command messages delivered to the server, then a pause message is sent to the client indicating which said command message cannot be received; indicia is stored that indicates the command message initially discarded; and subsequent said command messages delivered to the server are discarded until an initially discarded said command message is again delivered to the server.

Claims

1. A method for controlling command message flow in a network including a server and a client comprising the steps of:

including a command window, comprising a maximum number of command messages that may be outstanding at the server, in messages sent from the server to the client;

modifying the value of the command window at the server in accordance with available server resources for receiving said command messages;

sending a plurality of said command messages from the client to the server;

when there are insufficient resources at the server to process one of the command messages delivered to the server, then;

sending a pause message to the client indicating which said command message cannot be received;

storing indicia indicating the command message initially discarded; and

discarding subsequent said command messages delivered to the server until an initially discarded said command message is again delivered to the server;

ceasing sending said command messages from the client when the number of outstanding said command messages is at least equal to the maximum number of messages indicated by the command window; and

when the number of outstanding said command messages is less than the maximum number indicated by the command window, then resuming sending said command messages from the client to the server, starting with the command message initially discarded by the server.

2. The method of claim 1, wherein, after ceasing sending said command messages from the client when the number of outstanding said command messages is at least equal to the maximum number of messages indicated by the command window, the sending of command messages from the client is resumed if the number of outstanding said command messages decreases below the maximum number.

3. The method of claim 1, wherein the command window indicates the maximum number of commands that may be outstanding at the server, for each connection between the client and the server, based upon the availability of resources needed to receive said command messages on a presently established connection, the method including the step of allowing the sum of a plurality of the command windows to exceed the totality of the resources available at the server.

4. The method of claim 1, wherein, when the server sends said pause message, if the command window is larger than a predefined minimum and larger than the number of commands currently outstanding, then the command window is set to the larger of the minimum or the number of commands currently outstanding.

5. The method of claim 1, including the steps of:

maintaining for each connection, at the server, a command window request counter, initially set to a predefined initial value;

marking as delayed, each said command message that cannot be transmitted by the client because the maximum number of said command messages, as indicated by the command window, are presently outstanding;

marking as retransmitted, each said command message that must be retransmitted due to receipt of said pause message;

including, in each said command message transmitted, an indication of whether the command message was marked as either delayed or retransmitted;

when one of the command messages is received at the server with the indication that it was marked as either delayed or retransmitted, then decrementing the command window request counter by one; and

when the command window request counter at the server reaches a value of zero, then resetting the request counter to the initial value; and

incrementing the command window value by one, if the command window value is less than a predefined upper limit for the maximum number of said command messages.

6. A method for controlling command message flow in a network including a server and a client comprising the steps of:

including a command window, comprising a maximum number of command messages that may be outstanding at the server, in messages sent from the server to the client;

modifying the value of the command window at the server in accordance with available server resources for receiving said command messages;

when there are insufficient resources at the server to process one of the command messages delivered to the server, then;

sending a pause message to the client indicating which said command message cannot be received;

storing indicia indicating the command message initially discarded; and

discarding subsequent said command messages delivered to the server until an initially discarded said command message is again delivered to the server.

7. The method of claim 6, including:

sending a plurality of said command messages from the client to the server;

ceasing sending said command messages from the client when the number of outstanding said command messages is at least equal to the maximum number of messages indicated by the command window; and

when the number of outstanding command messages is less than the maximum number indicated by the command window, then resuming sending said command messages from the client to the server, starting with the command message initially discarded by the server.

8. The method of claim 6, including the steps of:

maintaining for each connection, at the server, a command window request counter, initially set to a predefined initial value;

marking as delayed, each said command message that cannot be transmitted by the client because the maximum number of command messages, as indicated by the command window, are presently outstanding;

marking as retransmitted, each said command message that must be retransmitted due to receipt of said pause message;

including, in each said command message transmitted, an indication of whether the command message was marked as either delayed or retransmitted;

when a specific said command message is received at the server with the indication that it was marked as either delayed or retransmitted, then decrementing the command window request counter by one; and

when the command window request counter at the server reaches a value of zero, then resetting the request counter to the initial value; and

incrementing the command window value by one, if the command window value is less than a predefined upper limit for the maximum number of said command messages.

9. The method of claim 6, wherein the command window indicates the maximum number of commands that may be outstanding at the server, for each connection between the client and the server, based upon the availability of resources needed to receive said command messages on a presently established connection, the method including the step of allowing the sum of a plurality of the command windows to exceed the totality of the resources available at the server.

10. The method of claim 6, wherein, after ceasing sending said command messages from the client when the number of outstanding said command messages is at least equal to the maximum number of messages indicated by the command window, the sending of command messages from the client is resumed if the number of outstanding said command messages decreases below the maximum number.

11. The method of claim 6, wherein, when the server sends said pause message, if the command window is larger than a predefined minimum and larger than the number of commands currently outstanding, then the command window is set to the larger of the minimum or the number of commands currently outstanding.

12. A method for controlling command message flow in a network including a server and a client comprising the steps of:

sending a plurality of command messages from the client to the server;

receiving, from the server, a message command window, comprising a maximum number of said command messages that may be outstanding at the server,

ceasing sending said command messages from the client when the number of outstanding command messages is at least equal to the maximum number of messages indicated by the command window; and

when the number of outstanding said command messages is less than the maximum number indicated by the command window, then resuming sending said command messages from the client to the server, starting with the command message initially discarded by the server.

13. The method of claim 12, wherein, after ceasing sending said command messages from the client when the number of outstanding said command messages is at least equal to the maximum number of messages indicated by the command window, the sending of command messages from the client is resumed if the number of outstanding said command messages decreases below the maximum number.

14. The method of claim 12, comprising the steps of:

including the command window in messages sent from the server to the client;

modifying the value of the command window at the server in accordance with available server resources for receiving said command messages;

when there are insufficient resources at the server to process one of the command messages delivered to the server, then;

sending a pause message to the client indicating which said command message cannot be received;

storing indicia indicating the command message initially discarded; and

discarding subsequent said command messages delivered to the server until an initially discarded said command message is again delivered to the server.

15. The method of claim 14, including the steps of:

maintaining for each connection, at the server, a command window request counter, initially set to a predefined value;

marking as delayed, each said command message that cannot be transmitted by the client because the maximum number of command messages, as indicated by the command window, are presently outstanding;

marking as retransmitted, each said command message that must be retransmitted due to receipt of a pause message;

including, in each said command message transmitted, an indication of whether the command message was marked as either delayed or retransmitted;

when one of the command messages is received at the server with the indication that it was marked as either delayed or retransmitted, then decrementing the command window request counter by one; and

when the command window request counter at the server reaches a value of zero, then resetting the request counter to the initial value; and

incrementing the command window value by one, if the command window value is less than a predefined upper limit for the maximum number of said command messages.

16. A method for controlling command message flow in a network including a server and a client comprising the steps of:

including a command window, comprising a maximum number of command messages that may be outstanding at the server, in each message sent from the server to the client;

setting an initial value of the command window via a message from the server to the client during connection establishment therebetween;

modifying the value of the command window at the server in accordance with available server resources for receiving said command messages;

sending a plurality of said command messages from the client to the server in client command message order until the number of outstanding command messages is equal to the maximum number indicated by the command window;

establishing command order at the server based on the order in which said command messages are delivered to the server by the sequenced message transport;

when there are insufficient resources at the server to process one of the command messages delivered to the server, then;

sending a pause message to the client indicating which said command message cannot be processed;

storing indicia indicating the command message initially discarded; and

discarding subsequent said command messages delivered to the server until the initially discarded said command message is again delivered to the server;

ceasing sending said command messages from the client when the number of outstanding command messages is at least equal to the maximum number of messages indicated by the command window;

when the number of outstanding command messages is less than the maximum number indicated by the command window, then resuming sending said command messages from the client to the server, starting with the command message initially discarded by the server, and proceeding to process received command messages in the client command message order; and

when the initially discarded said command message is again delivered to the server, then resuming processing of delivered command messages if sufficient server resources are available.

17. The method of claim 16, including the steps of:

maintaining for each connection, at the server, a command window request counter, initially set to a predefined initial value;

marking as delayed, each said command message that cannot be transmitted by the client because the maximum number of said command messages, as indicated by the command window, are presently outstanding;

marking as retransmitted, each said command message that must be retransmitted due to receipt of said pause message;

including, in each said command message transmitted, an indication of whether the command message was marked as either delayed or retransmitted;

when one of the command messages is received at the server with the indication that it was marked as either delayed or retransmitted, then decrementing the command window request counter by one; and

when the command window request counter at the server reaches a value of zero, then resetting the request counter to the initial value; and

incrementing the command window value by one, if the command window value is less than a predefined upper limit for the maximum number of said command messages.

18. The method of claim 16, wherein the command window indicates the maximum number of commands that may be outstanding at the server, for each connection between the client and the server, based upon the availability of resources needed to receive said command messages on a presently established connection, the method including the step of allowing the sum of a plurality of the command windows to exceed the totality of the resources available at the server.

19. A system for controlling command message flow in a network including a server and a client comprising:

a command window, comprising a maximum number of command messages that may be outstanding at the server, included in each message sent from the server to the client;

wherein;

an initial value of the command window is set via a message from the server to the client during connection establishment therebetween;

the value of the command window at the server is modified in accordance with available server resources for receiving said command messages;

a plurality of said command messages are sent from the client to the server in client command message order until the number of outstanding command messages is equal to the maximum number indicated by the command window;

command order at the server is established based on the order in which said command messages are delivered to the server by the sequenced message transport;

when there are insufficient resources at the server to process one of the command messages delivered to the server, then;

a pause message is sent to the client indicating which said command message cannot be processed;

indicia is stored indicating the command message initially discarded; and

subsequent said command messages delivered to the server are discarded until the initially discarded said command message is again delivered to the server;

said command messages from the client cease being sent to the server when the number of outstanding command messages is at least equal to the maximum number of messages indicated by the command window;

when the number of outstanding command messages is less than the maximum number indicated by the command window, then resuming sending said command messages from the client to the server, starting with the command message initially discarded by the server, and proceeding to process received command messages in the client command message order; and

when the initially discarded said command message is again delivered to the server, then processing of delivered command messages is resumed if sufficient server resources are available.

20. The system of claim 19, including:

a command window request counter, maintained at the server, for each connection, and initially set to a predefined value;

wherein;

each said command message that cannot be transmitted by the client because the maximum number of command messages, as indicated by the command window, are presently outstanding, is marked as delayed;

each said command message that must be retransmitted due to receipt of a pause message is marked as retransmitted;

an indication of whether the command message was marked as either delayed or retransmitted is included in each said command message transmitted;

when one of the command messages is received at the server with the indication that it was marked as either delayed or retransmitted, then the command window request counter is decremented by one; and

when the command window request counter at the server reaches a value of zero, then the request counter is reset to the initial value; and

the command window value is incremented by one, if the command window value is less than the maximum number of command messages.