Proxy server, hierarchical network system, and distributed workload management method

• US 9,712,448 B2
• Filed: 03/27/2012
• Issued: 07/18/2017
• Est. Priority Date: 03/30/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Wu, Hai Shan, Zhao, Yi Xin, Wang, Wen Jie, Yang, Bo
• Primary Examiner(s): Christensen, Scott B
• Assistant Examiner(s): Do, Lam
• Application Number:
  US13/430,948
• Publication Number:
  US 20120254413A1
• Time in Patent Office: 1,939 Days
• Field of Search: None
• CPC Class Codes: H04L 47/22 Traffic shaping
• CPC Class Codes: H04L 47/2425 for supporting services specification, e.g. SLA
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 67/51 Discovery or management thereof, e.g. service location protocol [SLP] or web services

Abstract

A proxy server and a hierarchical network system and a distributed workload management method. According to one embodiment of this disclosure, the proxy server includes: a rate controller configured to, based on measured request-related information and service quality parameters relative to service levels of requests, periodically determine a dispatch rate for requests of each service level, wherein the sum of the dispatch rate for respective service levels is less than or equal to a predetermined rate; and a request dispatcher configured to dispatch the requests of the corresponding service level in accordance with the dispatch rate determined by the rate controller. One aspect of the disclosure realizes a low overhead, highly scalable, simple and efficient workload management system to achieve QoS assurance and overload protection.

Claims

1. A method for distributed workload management in hierarchical network system, comprising:

maintaining, by a proxy server, a plurality of queues each corresponding to a different service level;

periodically determining, by the proxy server, a dispatch rate for requests for a current processing cycle of each of the service levels based on measured request-related information and service quality parameters relative to service levels of requests, wherein the measured request-related information is measured by the proxy server for a previous processing cycle, the periodically determining comprising;

associating, by the proxy server, the measured request-related information with the service quality parameters for the respective service levels according to a utility function, wherein the utility function represents a loss due to unsatisfied target response times of the respective service levels and is specified in terms of target response time and an importance of the requests;

calculating, by the proxy server, the dispatch rate for the corresponding service level by minimizing the utility function, wherein a sum of said dispatch rate for respective service levels is less than or equal to a predetermined rate specifying a maximum admitted rate to the proxy server; and

dispatching, by the proxy server, the requests of said corresponding service level in accordance with said determined dispatch rate to the plurality of application servers.

2. The method according to claim 1, wherein said measured request-related information is a current request response time, and said service quality parameters comprise a target response time and an importance of said requests of the corresponding service level.

3. The method according to claim 1, wherein periodically determining a dispatch rate for requests of each service level comprises:

correlating said measured request-related information and said service quality parameters relative to service levels of requests by using a utility function;

correlating the measured request-related information and the dispatch rate through a queuing model.

4. The method according to claim 1, wherein said predetermined rate is determined by an application server serving for a proxy server in said hierarchical network system, independent of said information about said proxy server.

5. The method according to claim 1, further comprising:

receiving requests from clients, and classifying said requests into one or more service levels; and

maintaining one or more request queues, wherein different queues correspond to different service levels of requests.

6. The method according to claim 1, wherein said predetermined rate is obtained through:

periodically detecting a current usage of a resource of said application server in said hierarchical network system and a current request arrival rate;

calculating a maximum allowable rate of requests for the next management cycle based on said detected current resource usage, said current request arrival rate and a target usage of the resource; and

allocating said maximum allowable rate to each proxy server in said hierarchical network system as its predetermined rate based on a predetermined policy.

7. The method according to claim 6, wherein said predetermined policy is to evenly allocate said maximum allowed rate to said proxy server.

8. The method according to claim 6, wherein said resources comprises at least one of CPU usage, memory usage, I/O, and bandwidth.

9. The method according to claim 8, wherein when said resource is CPU usage, said maximum rate calculator adopts a proportional-integral controller technology to calculate said maximum allowed rate.

10. The method according to claim 6, wherein the calculation period of said proxy server is in synchronization with said management cycle of the application server.