Demand responsive method and apparatus to automatically activate spare servers
View Patent ↗A server and method of its operation adapt the number of server applications within the server. The server is connected to a computer network. The server comprises one or more active server applications, a load detector, an inactive additional server application and an allocator. The load detector, which may be part of a load balancer, is connected to the one or more server applications and the computer network. The allocator is connected to the load detector and the additional server application. The allocator causes the additional server application to activate in response to a load condition. The method measures a load on the server, detects when the load exceeds a threshold and, in response thereto, activates an additional server application on the server. Optionally, the method also detects when the load is less than a deactivation threshold, and in response thereto, deactivates the additional server application. Also disclosed is a system comprising a plurality of computers, one or more connections to one or more servers, and a module. Each of the computers is capable of hosting a server application. The module is connected to the: computers and the connection(s). The module receives a request for an additional server application from one of the servers. Such a request may, for example, be generated when the server is experiencing a surge. In response to the request, the module activates the server application on one or more of the computers so as to support the requesting server.
1. A server connected to a computer network, the server comprising:
one or more active server applications;
a load detector connected to the one or more server applications and the computer network;
an additional server application that is inactive; and
an allocator connected to the load detector and the additional server application;
wherein the allocator causes the additional server application to activate in response to a load condition and wherein an activation threshold is based at least in part on a ratio of a current number of connections per unit time to a past number of connections per unit time.
2. The server of claim 1 further comprising: a load balancer connected to the load detector.
3. The server of claim 1 wherein each server application runs on a separate computer.
4. The server of claim 1 wherein the additional server application is in a power-on condition.
5. The server of claim 1 wherein the additional server application is owned by a computer vendor.
6. The server of claim 1 wherein the additional server is remotely located from the one or more server applications.
7. The server of claim 6 wherein the allocator comprises a modem connected to the location of the additional server.
8. The server of claim 1 wherein the allocator comprises a load threshold comparator.
9. The server of claim 1 wherein the allocator comprises a log file in which the activation of the additional server is logged.
10. A method of adapting the number of server applications within a server, the method comprising:
measuring a load on the server;
detecting when the load exceeds an activation threshold; and
in response to the detecting step, employing a system allocator to activate an additional server application on the server;
wherein the additional server application is connected to the server via the system allocator and wherein an activation threshold is based at least in part on a ratio of a current number of connections per unit time to a past number of connections per unit time.
11. The method of claim 10 wherein the load measuring step comprises one or more of the group selected from measuring CPU utilization, measuring a percentage of session usage, and measuring the rate of change of demand.
12. The method of claim 10 wherein the activation threshold is based on the quantities R/(C*M) and (R−r)/r wherein R is a current number of connections per unit time, C 1 is a maximum number of connections per unit time, M is the number of server applications currently activated, and r is a past number of connections per unit time.
13. The method of claim 10 wherein the activating step comprises:
loading an operating system on the additional server application; and configuring the operating system.
14. The method of claim 10 wherein the activating step comprises:
electronically contacting a computer vendor to request the additional server application.
15. The method of claim 10 wherein the activating step comprises:
using the additional server to accommodate a surge.
16. The method of claim 10 further comprising: logging the activating step.
17. The method of claim 10 further comprising:
detecting when the load is less than a deactivation threshold; and
in response to the minimum detecting step, deactivating the additional server application.
18. A system comprising:
a plurality of computers, each computer capable of hosting a server application;
one or more connections to one or more servers; and
a module connected to the one or more connections and the plurality of computers;
wherein the module receives a request for an additional server application from one of the plurality of servers and, in response, activates the server application on one or more of the plurality of computers so as to support the requesting server;
wherein said server application is connected to the requesting server via the module and wherein an activation threshold is based at least in part on a ratio of a current number of connections per unit time to a past number of connections per unit time.
19. The system of claim 18 wherein at least some of the one or more servers are Internet website servers.
20. The system of claim 18 further comprising a computer usage log, whereby usage-based billing can be derived.