Response time estimation for intermittently-available nodes

• US 20080298235A1
• Filed: 05/30/2007
• Published: 12/04/2008
• Est. Priority Date: 05/30/2007
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Neugebauer, Mario
• Granted Patent: US 7,855,975 B2
• Time in Patent Office: Days
• US Class Current: 370/232
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: H04L 43/0864 Round trip delays

Abstract

Message exchanges may be executed between a middleware component and at least one intermittently-available node, during connected phases of the intermittently available node with the middleware component. Transmission parameters associated with the message exchanges may be monitored, and a current request message to be sent to the intermittently-available node may be determined, the current request message being associated with a time-to-live. An expected message exchange time of the current request message and a resulting response message received from the node in response thereto may be determined, relative to the connected phases and based on at least some of the transmission parameters. Based on the expected message exchange time, it may be predicted whether the current request message and the resulting response message will be successfully exchanged with the intermittently-available node, prior to an expiration of the time-to-live of the current request message.

Claims

1. A response time estimator system, the system comprising:

a parameter monitor configured to monitor transmission parameters associated with message exchanges between a middleware component and at least one intermittently-available node, the intermittently-available node experiencing connected phases and non-connected phases with the middleware component;

an expectation manager configured to determine a current request message at a queue of the middleware component, and configured to determine an expected message exchange time for the current request message and a resulting response message, based on a number of preceding request messages within the queue and on corresponding response messages from the intermittently-available node, and based on at least some of the transmission parameters; and

a result predictor configured to perform a comparison of the expected message exchange time and a time-to-live of the request message, and configured to predict, based on the comparison, whether a receipt of the resulting response message will occur, within one of the connected phases and before expiration of the time-to-live.

2. The system of claim 1 wherein the parameter monitor includes an availability monitor that is configured to monitor characteristics of the connected phases of the intermittently-available node.

3. The system of claim 1 wherein the parameter monitor includes a data size monitor that is configured to monitor data sizes of the request messages and the resulting response messages.

4. The system of claim 1 wherein the parameter monitor includes a processing time monitor that is configured to determine a processing time associated with transmission of each of the request messages and corresponding receipt of each of the corresponding response messages.

5. The system of claim 1 wherein the intermittently-available node is associated with a moving or movable object.

6. The system of claim 1 wherein the intermittently-available node experiences the non-connected phases during a power save operation of the intermittently-available node.

7. The system of claim 1 comprising an expectation manager configured to determine an expected time of transmission of the current request message and the preceding request messages, and further configured to determine an expected time of transmission of the resulting response message and the corresponding response messages.

8. The system of claim 1 comprising a result predictor configured to determine one of the connected phases in which the resulting response message would be received.

9. The system of claim 1 comprising a result predictor configured to predict whether the receipt of the resulting response message will occur, prior to an expiration of the time-to-live.

10. The system of claim 1 comprising a result predictor configured to make a determination that the resulting response message will not be received prior to an end of the time-to-live.

11. The system of claim 10 wherein at least one action is taken in response to the determination, the at least one action including removal of the request message from the queue, increasing of the time-to-live, re-scheduling of the transmission of the request message, or outputting of a notification of increased latency.

12. A method comprising:

executing message exchanges between a middleware component and at least one intermittently-available node, during connected phases of the intermittently available node with the middleware component;

monitoring transmission parameters associated with the message exchanges;

determining a current request message to be sent to the intermittently-available node, the current request message being associated with a time-to-live;

determining an expected message exchange time of the current request message and a resulting response message received from the node in response thereto, relative to the connected phases and based on at least some of the transmission parameters; and

predicting, based on the expected message exchange time, whether the current request message and the resulting response message will be successfully exchanged with the intermittently-available node, prior to an expiration of the time-to-live of the current request message.

13. The method of claim 12 wherein monitoring transmission parameters associated with the message exchanges comprises:

monitoring one or more of availability parameters associated with characteristics of the connected phases of the intermittently-available node, data sizes of messages in the message exchanges, and processing times of pairs of the message exchanges.

14. The method of claim 12 wherein determining a current request message to be sent to the intermittently-available node, the current request message being associated with a time-to-live, comprises:

determining placement of the current request message into a queue.

15. The method of claim 14 wherein determining an expected message exchange time of the current request message and a resulting response message received from the node in response thereto, relative to the connected phases and based on at least some of the transmission parameters, comprises:

determining a number of preceding request messages in the queue; and

determining the expected message exchange time based on the number of preceding requests.

16. The method of claim 14 wherein determining an expected message exchange time of the current request message and a resulting response message received from the node in response thereto, relative to the connected phases and based on at least some of the transmission parameters, comprises:

determining a connected phase in which the resulting response message is expected to be received.

17. The method of claim 12 wherein predicting, based on the expected message exchange time, whether the current request message and the resulting response message will be successfully exchanged with the intermittently-available node, prior to an expiration of the time-to-live of the current request message, comprises:

predicting that the current request message and the resulting response message will not be successfully exchanged with the intermittently-available node, prior to an expiration of the time-to-live.

18. A message exchange system comprising

a queue configured to execute message exchanges with at least one intermittently-available node, during connected phases of the intermittently available node;

a message handler configured to forward request messages from the queue to the intermittently-available node, based in part on a time-to-live associated with each message, and configured to receive resulting response messages received from the intermittently-available node; and

a response time estimator configured to make a comparison of the time-to live of a current request message in the queue with an expected message exchange time of the current request message and a resulting response message from the intermittently-available node, and configured to predict a result of a future exchange of the current request message with the resulting response message.

19. The message exchange system of claim 18 wherein the response time estimator is configured to determine a future failure of the future exchange request, prior to an expiration of the time-to-live.

20. The message exchange system of claim 19, comprising one or more of a task scheduler configured to re-schedule transmission of the current request message, based on the future failure, a latency notifier configured to output an indication of increased latency of the message exchange system based on the future failure, and a time-to-live (TTL) optimizer configured to adjust a value of the time-to-live, based on the future failure.