Service placement for enforcing performance and availability levels in a multi-node system

• US 20050038829A1
• Filed: 08/12/2004
• Published: 02/17/2005
• Est. Priority Date: 08/14/2003
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Chidambaran, Lakshminarayanan, Kantarjiev, Christopher A.
• Granted Patent: US 7,437,460 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• 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]

Abstract

An approach efficiently and dynamically places services within a multi-node system when expanding or contracting services, that is, increasing and decreasing the number of instances that host a service. Service placement decisions are made in a way that accounts for performance and availability requirements of both the service being placed and other services.

Claims

1. A method for managing resources in a multiple node system, the method comprising the steps of:

a plurality of servers hosting a service of a plurality of services;

wherein each service of said plurality of services is performance of work of a particular type;

detecting a violation of a service-level agreement of a plurality of service-level agreements;

in response to detecting the violation, determining whether a first service of said plurality of services may be hosted by a first server of said plurality of servers that is not hosting said first service; and

if said first service may be hosted by said first server, then causing said first server to host said first service.

2. The method of claim 1, the steps further including:

selecting said first service based on a performance measure associated with each service of said plurality of services; and

wherein selecting at least one service causes said determining whether a first service of said plurality of services may be hosted by a first server of said plurality of services not already hosting said first service.

3. The method of claim 2, wherein said performance measure is based on CPU utilization by said each service.

4. The method of claim 2, wherein said performance measure is based on a number of servers hosting said service.

5. The method of claim 1, wherein selecting at least one server of said plurality of servers includes selecting at least one server of said plurality of servers based on a performance measure associated with each server of said plurality of servers.

6. The method of claim 5, the steps further including causing a service-level agreement to be ignored when selecting at least one server of said plurality of servers based on a performance measure.

7. The method of claim 5, wherein selecting at least one server of said plurality of servers based on a performance measure includes selecting another server based on data that indicates that the first service performs better when co-located with a particular service on said other server relative to other services.

8. The method of claim 5, wherein said performance measure is based on CPU utilization of said each server.

9. The method of claim 1, wherein:

the steps further include, in response to detecting the violation, determining not to perform session balancing between at least one of said plurality of services and another of said plurality of servers; and

the step of determining whether a first service of said plurality of services may be hosted is performed in response to determining not to perform session balancing.

10. The method of claim 9, wherein the steps further include after performing the step of causing said first server to host said first service:

determining that said violation persists; and

determining whether to perform session balancing between at least one of said plurality of services and another of said plurality of servers.

11. The method of claim 1, wherein the steps further include:

after performing the step causing said first server to host said first service, determining that said violation persists; and

wherein determining that said violation persists causes sessions to be migrated from another server of said plurality of servers to said first server.

12. The method of claim 1, wherein each server of said plurality of servers manages access to a database.

13. The method of claim 1, wherein:

the plurality of servers execute on a set of nodes with each node having shared access to a non-volatile storage; and

each node of said set of nodes accesses a data item stored on said non-volatile storage without requiring another node access said data item on said each node'"'"'s behalf.

14. The method of claim 2, wherein selecting said first service includes invoking a user-supplied routine that returns a certain server to host said first service.

15. A method for managing resources in a multiple node system, the method comprising the steps of:

a plurality of servers hosting a service of a plurality of services;

wherein each service of said plurality of services is performance of work of a particular type;

detecting a violation of a service-level agreement of a plurality of service-level agreements;

in response to detecting the violation, selecting a server of said plurality of servers to cease hosting a first service of said plurality of services;

wherein said selecting the server is based on one or more factors that indicate the cost of causing said server to cease hosting said first service; and

in response to selecting the server, causing said server to cease to host said service.

16. The method of claim 15, wherein the one or more factors are based on one or more performance measures.

17. The method of claim 16, wherein the one or more performance measures include CPU utilization for the service.

18. The method of claim 16, wherein the one or more performance measures is based on a minimum cardinality established for said first service.

19. The method of claim 16, wherein the one or more factors include a number of sessions on said server that is associated with said first service.

20. The method of claim 16, wherein the step of selecting a server of said plurality of servers to cease hosting a first service includes invoking a user-supplied routine that returns a certain server to cease hosting said first service.