Methods and systems for dynamic division of path capacity
View Patent ↗Methods and systems are disclosed that relate to selecting a path for sending an I/O request from a host to a data storage subsystem from among a plurality of paths from the host to the subsystem. An exemplary method includes identifying a limitation on the traffic level for the plurality of paths, tracking a first metric corresponding to the limitation on the traffic level for each path, and transmitting a first I/O request having an urgency level other than the highest urgency level by one of the plurality of paths whose first metric does not exceed its limitation on the traffic level.
1. A method for selecting a path by which to send an I/O request from a host to an associated storage subsystem from among a plurality of paths from the host to the associated storage subsystem, the method comprising:
identifying, for each of the plurality of paths, a first limitation on a traffic level for I/O requests having a first urgency level, wherein the first urgency level does not comprise the highest urgency level;
tracking a first metric corresponding to the first limitation for each of the plurality of paths;
identifying a first set of paths whose first metric does not exceed its first limitation; and
transmitting a first I/O request having the first urgency level by one of the first set of identified paths;
associating the first I/O request with one of a plurality of I/O groups, each I/O group comprising I/O requests sharing a distinguishing characteristic;
identifying a performance target for each I/O group;
tracking, for the first I/O request, a metric corresponding to the performance target for the associated I/O group;
assigning an urgency level to the first I/O request based on the difference between the performance target and the metric,
wherein when the metric is approximately equal to or greater than the performance target, the highest urgency level is assigned to the first I/O request.
2. The method of claim 1 , further comprising transmitting a second I/O request having the highest urgency level by the path having the lowest traffic level among the plurality of paths.
3. The method of claim 1 , further comprising transmitting the first I/O request by one of the first set of identified paths having the lowest traffic level among the plurality of paths.
4. The method of claim 1 , further comprising:
identifying, for each of the plurality of paths, a second limitation on the traffic level for I/O requests having a second urgency level, wherein the second urgency level does not comprise the highest urgency level;
tracking a second metric corresponding to the second limitation for each of the plurality of paths;
identifying a second set of paths whose second metric does not exceed its second limitation; and
transmitting a second I/O request having the second urgency level by one of the second set of identified paths.
5. The method of claim 1 , wherein the limitation on the traffic level comprises a number of pending I/O requests.
6. The method of claim 1 , wherein the first limitation on the traffic level comprises a number of bytes in pending I/O requests.
7. The method of claim 1 , wherein the first limitation on the traffic level comprises a time between transmissions of I/O requests corresponding to the first urgency level along the path.
8. The method of claim 1 , further comprising:
identifying, for each of the plurality of paths, a second limitation on the traffic level for I/O requests corresponding to the first urgency level;
tracking a second metric corresponding to the second limitation for each of the plurality of paths; and
transmitting a first I/O request having the first urgency level by one of identified paths, wherein the first set of paths comprises paths whose second metric does not exceed its second limitation.
9. The method of claim 8 , wherein the first limitation and the second limitation are selected from the group consisting of a number of pending I/O requests, a number of bytes in pending I/O requests, and a time between sending non-urgent I/O requests.
10. The method of claim 1 , further comprising queuing I/O requests having a first urgency level on the host when the first metric for each path exceeds its first limitation.
11. The method of claim 1 , wherein when the metric is less than the performance target, an urgency level other than the highest urgency level is assigned to the first I/O request.
12. The method of claim 11 , wherein the assigning is periodically performed and the assigned urgency level is increased as the metric approaches the performance target.
13. The method of claim 1 , wherein the distinguishing characteristic is selected from the group consisting of an I/O type, an I/O size, a host application initiating the I/O request, and a logical unit to which the I/O request is directed.
14. The method of claim 1 , wherein the performance target is a maximum allowable delay before the I/O request is sent to an associated storage subsystem.
15. The method of claim 1 , wherein the performance target is a minimum number of I/O requests per second sent to an associated storage subsystem.
16. A method for selecting a path by which to send an I/O request from a host to an associated storage subsystem from among a plurality of paths from the host to the associated storage subsystem, the method comprising:
identifying, for each of the plurality of paths, a maximum number of pending I/O requests having a first urgency level, wherein the first urgency level does not comprise the highest urgency level;
tracking a number of pending I/O requests having the first urgency level for each of the plurality of paths;
identifying a first set of paths whose number of pending I/O requests having the first urgency level does not exceed its identified maximum number; and
transmitting a first I/O request having the first urgency level by one of the first set of identified paths;
associating the first I/O request with one of a plurality of I/O groups, each I/O group comprising I/O requests sharing a distinguishing characteristic;
identifying a performance target for each I/O group;
tracking, for the first I/O request, a metric corresponding to the performance target for the associated I/O group;
assigning an urgency level to the first I/O request based on the difference between the performance target and the metric,
wherein when the metric is approximately equal to or greater than the performance target, the highest urgency level is assigned to the first I/O request.
17. The method of claim 16 wherein, further comprising:
identifying, for each of the plurality of paths, a second limitation on the traffic level for I/O requests having a second urgency level, wherein the second urgency level does not comprise the highest urgency level;
tracking a second metric corresponding to the second limitation for each of the plurality of paths;
identifying a second set of paths whose second metric does not exceed its second limitation; and
transmitting a second I/O request having the second urgency level by one of the second set of identified paths.
18. A method, comprising:
selecting a path by which to send an I/O request from a host to an associated storage subsystem from among a plurality of paths from the host to the associated storage subsystem, the selecting comprising:
identifying, for each of the plurality of paths, a first traffic level limitation for I/O requests having a first urgency level, wherein the first urgency level does not comprise the highest urgency level;
tracking a first metric corresponding to the first limitation for each of the plurality of paths;
identifying a first set of paths whose first metric does not exceed its identified first traffic level limitation; and
transmitting a first I/O request having the first urgency level by one of the first set of identified paths;
identifying a path among the plurality of paths having a lowest traffic level; and
transmitting a second I/O request having the highest urgency level by the path among the plurality of paths having the lowest traffic level;
associating the first I/O request with one of a plurality of I/O groups, each I/O group comprising I/O requests sharing a distinguishing characteristic;
identifying a performance target for each I/O group;
tracking, for the first I/O request, a metric corresponding to the performance target for the associated I/O group;
assigning an urgency level to the first I/O request based on the difference between the performance target and the metric,
wherein when the metric is approximately equal to or greater than the performance target, the highest urgency level is assigned to the first I/O request.
19. The method of claim 18 , wherein the method is performed by an I/O driver on the host.
20. The method of claim 18 wherein, further comprising:
identifying, for each of the plurality of paths, a second limitation on the traffic level for I/O requests having a second urgency level, wherein the second urgency level does not comprise the highest urgency level;
tracking a second metric corresponding to the second limitation for each of the plurality of paths;
identifying a second set of paths whose second metric does not exceed its second limitation; and
transmitting a second I/O request having the second urgency level by one of the second set of identified paths.