Method and apparatus for implementing a service-level agreement

• US 7,058,704 B1
• Filed: 11/30/1999
• Issued: 06/06/2006
• Est. Priority Date: 12/01/1998
• Status: Expired due to Fees

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Mangipudi, Krishna, Basani, Vijay
• Primary Examiner(s): Najjar, Saleh
• Assistant Examiner(s): NGUYEN, THU HA T
• Application Number:
  US09/451,592
• Time in Patent Office: 2,380 Days
• Field of Search: 709/223, 709/224, 709225-226, 709/232, 709/235, 709200-203, 370/270, 370/235, 370252-254
• US Class Current: 709/223
• CPC Class Codes: H04L 41/5009 Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
• CPC Class Codes: H04L 41/5019 Ensuring fulfilment of SLA
• CPC Class Codes: H04L 41/5032 Generating service level reports

Abstract

A method and apparatus is provided for generating, collecting, and manipulating useful information for validating or defining SLAs of web servers on a network. Web servers comprising a web farm on the network are adapted for logging detailed runtime information regarding user transactions and performance parameters. An Accumulator device interacts with intelligent agents to collect and combine their log files, process the combined file and post information into a database. An operator enters committed performance parameters into an SLA Reporter system according to classes of users, classes of web sites being hosted on the web servers, classes of URLs, transaction, content and file type. When compared with the database, processing of SLA reports indicate how well the parameters of the SLAs are being met for users, web sites, classes, URL'"'"'s and transactions, or other measurable elements. By generating, collecting, combining and processing in this manner, application-specific performance can be quickly and automatically evaluated with respect to parameters related to user satisfaction and detailed signals can be issued for cases in which remedial steps should be undertaken.

Claims

1. A method for evaluating compliance with a service-level agreement, said method comprising the steps of:

providing at least one back-end server connected to a network;

defining service level classes of service as a function of agreed-to service level metrics;

grouping at least one of host system, user, URL, hosted site, transaction, content and file type into respective service level classes of service;

defining said service level metrics to be measured for each of said classes;

defining acceptance levels for each class for each of said service level metrics;

collecting information related to measurement of said service level metrics;

comparing said acceptance levels to said information;

deploying an intelligent agent on each of said back end servers;

monitoring service level metrics using said intelligent agent and logging them into respective log files;

collecting log files from said intelligent agents on a list of said back-end servers; and

consolidating said log files collected from said list of back end servers in a database.

2. The method of claim 1 in which said step of defining further includes the steps of:

providing a format in which a set of servers will provide information to be measured; and

collecting said information.

3. The method of claim 1 in which said step of defining acceptance levels comprises the step of generating a database entry for each service commitment element of a service-level agreement.

4. The method of claim 1 in which the service level metrics are selected from the set consisting of:

records of performance, and errors, client IP address, username, date, time, service, server name, server IP address, processing time, bytes sent, bytes received, service status, operation, target URL, User Agent, referrer parameters, SMserver, Smvirtual site, and cookie.

5. The method of claim 1 wherein the information collected further includes information selected from the group consisting of:

assigned disk space, what the user can access, how the user'"'"'s request is fulfilled within the system or web farm, user'"'"'s subscribed level of service or class, transaction, number of requests, download size, file size, file type, time of day, week or month, response time of the back end servers, response time of the web farm, and how long it takes to complete a specified request or file.

6. The method of claim 1 in which said information related to measurement of said service level metrics is collected in real-time.

7. The method of claim 1 in wherein each of said host system, user, URL, hosted site, transaction, content and file type belong to no more than one class.

8. The method of claim 1 in which said step of defining service level classes of service comprises the steps of:

identifying a set of agreed-to metrics; and

associating said set of agreed-to metrics with one of a plurality of said service level classes of service.

9. The method of claim 1 wherein each of a plurality of said service level classes of service is defined by a set of agreed-to metrics.

10. The method of claim 1 wherein said service level classes of service define a hierarchy of service levels.

11. An apparatus for evaluating real-time compliance with a service-level agreement comprising:

at least one back-end server;

a network connecting said at least one back-end server;

a collection processor measuring and periodically collecting a set of defined service level metrics for said at least one back-end server;

a set of acceptance levels for said collected service level metrics wherein said set of acceptance levels define a service level class of service for at least one of host system, URL, hosted site, transaction, content, file type and user;

a monitoring processor determining which of said collected service level metrics exceed a corresponding acceptance level;

a reporting process that produces a report on a per-class basis of the results of said monitoring processor;

an intelligent agent deployed on each of said back-end servers monitoring said service level metrics and logging them into respective log files;

a scheduler triggering said reporter to begin collection of log files from a list of back-end servers;

an accumulator requesting log files from the intelligent agent of each listed back-end server and consolidating the log files into a database; and

an interface mechanism between said accumulator and each intelligent agent, said interface mechanism ensuring that each requested log file is completely transferred to the accumulator prior to starting consolidation.

12. The apparatus of claim 11 wherein each of said intelligent agents further keeps track of which portions of said log files have been transferred.

13. The apparatus of claim 11 wherein said service level class of service for at least one of host system, URL, hosted site, transaction, content, file type and user is defined by identifying a set of agreed-to metrics and associating said set of agreed-to metrics with one of a plurality of said service level classes of service.

14. A method for implementing management of a service level agreement monitoring system for a set of back-end servers in a web farm comprising the steps of:

providing at least one back-end server connected to a network;

defining a hierarchy of service level classes of service as a function of agreed-to service level metrics;

defining acceptance levels of said agreed to service level metrics for said service level classes;

selecting at least one class of service level metrics to be monitored;

creating a database of monitored service level metrics; and

preparing reports and/or alarms according to said selected class(s) of service level metrics;

deploying an intelligent agent on each of said back end servers;

monitoring service level metrics using said intelligent agent and logging them into respective log files;

collecting log files from said intelligent agents on a list of said back-end servers; and

consolidating said log files collected from said list of back end servers in said database.

15. The method of claim 14 further including the step of providing a graphical user interface for performing at least one of selecting the class of service level metrics to be monitored, defining thresholds of service-level commitments for at least some of said service level metrics, defining alarm trigger events, scheduling monitoring and reporting functions, and determining reporting formats.

16. The method of claim 15 wherein said step of preparing reports and/or alarms is further based on thresholds, schedules, and formats defined by the graphical user interface.

17. The method of claim 15 wherein the step of defining a hierarchy of service level classes is based on one of users, URLs and virtual sites.

18. The method of claim 14 in which said service level metrics to be monitored are selected from the set consisting of:

records of performance, transactions, and errors, client IP address, username, date, time, service, server name, server IP address, processing time, bytes sent, bytes received, service status, operation, target URL, User Agent, referrer parameters, SMserver, Smvirtual site, and cookie.

19. The method of claim 14 wherein said database of monitored service level metrics further includes information selected from the group consisting of:

the assigned disk space, what the user can access, how the user'"'"'s request is fulfilled within the system or web farm, user'"'"'s subscribed level of service or class, transaction, number of requests, download size, file size, file type, time of day, week or month, response time of the back end servers, response time of the web farm, and how long it takes to complete a specified request or file.

20. The method of claim 14 in which said classes of service level metrics are selected according to user class, host class, and virtual site class.

21. The method of claim 14 wherein the step of defining a hierarchy of service level classes is based on one of users, URLs and virtual sites.

22. The method of claim 14 wherein the step of defining service level classes of service comprises identifying a set of agreed-to metrics and associating said set of agreed-to metrics with one of a plurality of said service level classes of service.