Controlling access of concurrent users of computer resources in a distributed system using an improved semaphore counting approach

• US 7,743,146 B2
• Filed: 12/19/2007
• Issued: 06/22/2010
• Est. Priority Date: 01/30/2001
• Status: Expired due to Term

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Sadjadi, Shahrokh
• Primary Examiner(s): BOUTAH, ALINA A
• Application Number:
  US11/960,376
• Publication Number:
  US 20080162481A1
• Time in Patent Office: 916 Days
• Field of Search: 709/225, 709208-209, 709210-212, 709/217, 709/219, 709/223, 707/8, 710/200, 711151-153, 711/168
• US Class Current: 709/225
• CPC Class Codes: G06F 2209/522 Manager
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/52 specially adapted for the location of the user terminal
• CPC Class Codes: H04L 9/40 Network security protocols
• CPC Class Codes: H04W 4/02 Services making use of location information
• CPC Class Codes: H04W 4/023 using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database

Abstract

A method of controlling concurrent users of a distributed resource on a network is disclosed. In one aspect, there are one or more local lock managers executing on corresponding hosts and cooperating as a distributed lock manager. The resource is limited to a maximum number of concurrent users. A user identification for each user is associated with one host. In response to a request associated with a particular user associated with a first host, a lock is requested from a first local lock manager process executing on the first host. A related method of handling a request for a count-limited resource includes receiving a request from a client process for the computer resource. If it is determined that the request exceeds a maximum count for the resource, then it is determined whether a current time is within a retry time period of the client'"'"'s first request. If it is determined the current time is within the retry time period, then it is automatically determined again whether the request exceeds the maximum count for the resource. If it is determined that the request does not exceed the maximum count, then the resource is provided to the client process.

Claims

1. An apparatus for controlling use by concurrent users of a distributed resource on a network, wherein use of the distributed resource is limited to a specified maximum number of concurrent users, the apparatus comprising:

a first host comprising logic encoded in one or more tangible medium for execution;

wherein the first host is one of a plurality of hosts that are operable to communicatively connect to the network that contains the distributed resource;

wherein the logic when executed is operable to;

provide a local lock manager process that executes on the first host;

associate a user identification for each user of a first set of the concurrent users with the first host; and

respond to a first request for the distributed resource associated with a first user having a first user identification associated with the first host by requesting a lock from the local lock manager process;

wherein the local lock manager process may grant a lock on the same resource as another local lock manager process executing on another host of the plurality of hosts at any time when a number of outstanding locks granted on the same resource is less than the specified maximum number of concurrent users.

2. The apparatus of claim 1, wherein the logic is further operable to:

respond to a second request for the distributed resource associated with a second user having a second user identification associated with a second host of the plurality of hosts by requesting a lock from a second local lock manager process that executes on the second host;

wherein the second local lock manager process may grant a lock on the same resource as another local lock manager process executing on another host of the plurality of hosts at any time when the number of outstanding locks granted on the same resource is less than the specified maximum number of concurrent users.

3. The apparatus of claim 1, wherein the logic is operable to associate the first user with the first host based on information indicating that the first user signs onto the network most frequently at a terminal node of the network that uses fewer intervening network devices to pass messages to the first host than to any other host of the plurality of hosts.

4. The apparatus of claim 1, wherein the logic is operable to associate the first user with the first host based on information indicating that the first user signs onto the network most frequently at a terminal node of the network that is geographically closer to the first host than to any other of the plurality of hosts.

5. The apparatus of claim 1, wherein:

the local lock manager process comprises a lock data structure associated with the distributed resource and the lock data structure includes a local resource maximum field; and

the logic is further operable to generate and store a value in the local resource maximum field such that a summation over values in the local resource maximum fields maintained by a plurality of local lock manager processes, which respectively execute on the plurality of hosts, does not exceed the maximum number of concurrent users.

6. The apparatus of claim 5, wherein a particular value in the local resource maximum field maintained by the local lock manager process is based on a number of users associated with the first host.

7. The apparatus of claim 5, wherein the logic is further operable to:

determine whether a number of outstanding locks granted by the local lock manager process is less than a particular value stored in the local resource maximum field; and

generate and return a lock object associated with the lock when the number of outstanding locks is less than the particular value.

8. The apparatus of claim 1, wherein a copy of the distributed resource resides on a particular host of the plurality of hosts.

9. The apparatus of claim 1, wherein a copy of the distributed resource resides on a computing device that is not among the plurality of hosts.

10. An apparatus controlling concurrent users of a distributed resource on a network, wherein the distributed resource is limited to a maximum number of concurrent users, the apparatus comprising:

one or more network interfaces that are operable to communicatively connect to the network that contains the distributed resource; and

logic encoded in one or more tangible medium for execution and when executed operable to;

receive a request for the distributed resource from a client process for a user having a user identification;

determine a home location associated with the user identification, the home location indicating a unique host among a plurality of hosts that execute a corresponding plurality of local lock manager processes of a distributed lock manager process;

wherein each of the plurality of local lock manager processes may concurrently grant a lock on the same resource;

send a request for a lock object for the distributed resource to a first local lock manager process of the plurality of local lock manager processes, the request including the home location;

receive the lock object for the distributed resource from a second local lock manager process executing on the unique host when a number of outstanding locks granted by the second local lock manager process is less than a value of a local resource maximum defined for the second local lock manager process; and

provide the client process access to the distributed resource only in response to receiving the lock object.

11. The apparatus of claim 10, wherein a first host on which the first local lock manager process executes is different from the unique host on which the second local lock manager process executes.

12. The apparatus of claim 10, wherein a first host on which the first local lock manager process is executing is in a local network with a computing device on which the client process is executing.

13. The apparatus of claim 10, wherein the logic is further operable to:

determine whether the request received from the client process sets a retry time period; and

when the request received from the client process sets the retry time period, automatically resend within the retry time period the request for the lock object to the first local lock manager process.

14. The apparatus of claim 13, wherein:

each local lock manager process of the distributed lock manager process is configured to store a corresponding local maximum number of users for the distributed resource in a lock data structure associated with said each local lock manager process; and

an aggregate of local maximum numbers in lock data structures associated with all lock manager processes in the distributed lock manager process for the distributed resource is no greater than the maximum number of concurrent users to which the distributed resource is limited.

15. An apparatus controlling concurrent users of a distributed resource on a network, wherein the distributed resource is limited to a maximum number of concurrent users, the apparatus comprising:

one or more network interfaces that are operable to communicatively connect to the network that contains the distributed resource; and

logic encoded in one or more tangible medium for execution and when executed operable to;

generate a value for a local resource maximum stored on each host of a plurality of hosts such that a summation over the plurality of hosts of the value for the local resource maximum yields an aggregate value that does not exceed the maximum number of concurrent users;

determine whether to increase a first value in a first resource maximum stored on a first host of the plurality of hosts;

when the first value in the first resource maximum is to be increased, then;

decrease by a particular amount a second value in a second resource maximum stored on a second host of the plurality of hosts, and

increase by the particular amount the first value in the first resource maximum stored on the first host; and

grant a lock for the distributed resource when a number of outstanding locks granted by a lock manager process executing on the first host is less than the value of the local resource maximum stored on the first host.

16. The apparatus of claim 15, wherein the logic operable to determine whether to increase the first value is further operable to determine the time of day and the geographic location served by the first host.

17. The apparatus of claim 15, wherein the logic operable to determine whether to increase the first value is further operable to determine whether a difference between the number of outstanding locks granted and the first value is less than a first predetermined threshold.

18. The apparatus of claim 15, wherein when the first value in the first resource maximum is to be increased, the logic is further operable to determine whether to decrease the second value based on a whether a difference between the number of outstanding locks granted and the second value is greater than a second predetermined threshold.

19. The apparatus of claim 18, wherein the particular amount is based on the second predetermined threshold.

20. The apparatus of claim 19, wherein the particular amount is no greater than the second predetermined threshold.