Dynamic server farms

• US 9,154,549 B2
• Filed: 10/27/2011
• Issued: 10/06/2015
• Est. Priority Date: 10/27/2011
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Sundararaj, Carthic
• Primary Examiner(s): Gee, Jason K.
• Assistant Examiner(s): BUCKNOR, OLANREWAJU J
• Application Number:
  US13/283,114
• Publication Number:
  US 20130111467A1
• Time in Patent Office: 1,440 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/5011 Pool
• CPC Class Codes: G06F 2209/5022 Workload threshold
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers

Abstract

Techniques are provided for automatically adding/removing servers to/from a server farm in response to monitored load demands. A load among a plurality of servers, resulting from traffic associated with an application hosted on the plurality of servers is balanced. The load on the plurality of servers is monitored and it is determined that the load on the plurality of servers exceeds a predetermined load threshold. In response to such a determination, one or more additional servers are automatically provisioned for use in hosting the application. The load is then balanced between the plurality of servers and the one or more additional servers.

Claims

1. A method comprising:

balancing, among a plurality of servers, a load resulting from processing of traffic associated with an application hosted on the plurality of servers;

monitoring the load on the plurality of servers that results from the processing of the traffic associated with the application hosted on the plurality of servers;

determining that the load on the plurality of servers has caused a number of connections to at least one of the plurality of servers to exceed a predetermined load threshold indicative of a predetermined percentage of a maximum number of connections supported by the at least one of the plurality of servers, wherein the predetermined load threshold is chosen to account for a time needed to add an additional server so that the added server is ready when needed;

automatically provisioning one or more additional servers for use in hosting the application in response to determining that the load on the plurality of servers exceeds the predetermined load threshold;

assigning a cost to each of the plurality of servers and one or more additional servers that varies depending on whether the server is a reserved server or an on-demand server;

balancing the load resulting from the traffic associated with the application between the plurality of servers and the one or more additional servers;

determining that the load on the plurality of servers and the one or more additional servers is below the predetermined load threshold; and

in response to determining that the load is below the predetermined load threshold, automatically disabling at least one of the plurality of servers or one of the one or more additional servers based on the assigned costs such that higher cost servers are disabled first.

2. The method of claim 1, wherein automatically provisioning one or more additional servers comprises:

automatically installing an operating system on the one or more additional servers;

automatically installing the application on the one or more additional servers; and

automatically configuring the application on the one or more additional servers for processing the traffic associated with the application.

3. The method of claim 1, wherein automatically provisioning one or more additional servers comprises:

generating a request from a load balancer to a resource manager to provision the one or more additional servers for use in hosting the application.

4. The method of claim 3, further comprising:

generating a request that identifies a template for use by the resource manager in provisioning the one or more additional servers.

5. The method of claim 1, wherein the assigning a cost includes assigning a cost to each server that varies further depending on whether the type of server is a reserved server, an on-demand server, or a spot server.

6. One or more non-transitory computer readable storage media encoded with software comprising computer executable instructions and when the software is executed at a load balancer it is operable to:

balance, among a plurality of servers, a load resulting from processing of traffic associated with an application hosted on the plurality of servers;

monitor the load on the plurality of servers that results from the processing of the traffic associated with the application hosted on the plurality of servers;

determine that the load on the plurality of servers has caused a number of connections to at least one of the plurality of servers to exceed a predetermined load threshold indicative of a predetermined percentage of a maximum number of connections supported by the at least one of the plurality of servers, wherein the predetermined load threshold is chosen to account for a time needed to add an additional server so that the added server is ready when needed;

automatically provision one or more additional servers for use in hosting the application in response to a determination that the load exceeds a predetermined load threshold;

assign a cost to each of the plurality of servers and one or more additional servers that varies depending on whether the server is a reserved server or an on-demand server;

balance the load resulting from the traffic associated with the application between the plurality of servers and the one or more additional servers;

determine that the load on the plurality of servers and the one or more additional servers is below the predetermined load threshold; and

in response to determining that the load is below the predetermined load threshold, automatically disable at least one of the plurality of servers or one of the one or more additional servers based on the assigned costs such that higher cost servers are disabled first.

7. The non-transitory computer readable media of claim 6, wherein the instructions operable to automatically provision one or more additional servers comprise instructions operable to:

automatically install an operating system on the one or more additional servers;

automatically install the application on the one or more additional servers; and

automatically configure the application on the one or more additional servers for processing the traffic associated with the application.

8. The non-transitory computer readable media of claim 6, wherein the instructions operable to automatically provision one or more additional servers comprise instructions operable to:

generate a request from the load balancer to a resource manager to provision the one or more additional servers for use in hosting the application.

9. The non-transitory computer readable media of claim 8, further comprising instructions operable to:

generate a request that identifies a template for use by the resource manager in provisioning the one or more additional servers.

10. The computer readable medium of claim 6, wherein instructions to assign a cost include instructions to assign a cost to each server that varies further depending on whether the type of server is a reserved server, an on-demand server, or a spot server.

11. An apparatus comprising:

a plurality of network interfaces configured to receive and forward data traffic;

a memory; and

a processor configured to;

balance, among a plurality of servers, a load resulting from processing of traffic associated with an application hosted on the plurality of servers, monitor the load on the plurality of servers,

determine that the load on the plurality of servers has caused a number of connections to at least one of the plurality of servers to exceed a predetermined load threshold indicative of a predetermined percentage of a maximum number of connections supported by the at least one of the plurality of servers,

automatically provision one or more additional servers for use in hosting the application in response to determining that the load on the plurality of servers exceeds a predetermined load threshold, wherein the predetermined load threshold is chosen to account for a time needed to add an additional server so that the added server is ready when needed,

assign a cost to each of the plurality of servers and one or more additional servers that varies depending on whether the server is a reserved server or an on-demand server,

balance the load resulting from the traffic associated with the application between the plurality of servers and the one or more additional servers,

determine that the load on the plurality of servers and the one or more additional servers is below the predetermined load threshold, and

in response to determining that the load is below the predetermined load threshold, automatically disable at least one of the plurality of servers or one of the one or more additional servers based on the assigned costs such that higher cost servers are disabled first.

12. The apparatus of claim 11, wherein to automatically provision one or more additional servers, the processor is configured to automatically install an operating system on the one or more additional servers, automatically install the application on the one or more additional servers, and automatically configure the application on the one or more additional servers for processing the data traffic.

13. The apparatus of claim 11, wherein to automatically provision one or more additional servers, the processor is configured generate a request to a resource manager to provision the one or more additional servers for use in hosting the application.

14. The apparatus of claim 13, wherein the processor is further configured to generate a request that identifies a template for use by the resource manager in provisioning the one or more additional servers.

15. The apparatus of claim 11, wherein the processor is configured to assign to each server that varies further depending on whether the type of server is a reserved server, an on-demand server, or a spot server.