Controlling the operation of server computers
View Patent ↗A plurality of server computers in a network that includes load balancing processes to enhance performance employs a non-performance related variable, such as power consumption, and modifies the load balancing processes in response to the reading of the non-performance variable. Such variables do not affect the response and performance as perceived by a browsing client but do provide other advantages with in the environment as a whole. The non-performance related variable is employed such that one or more of said server computers are de-activated to reduce power consumption, and the load balancing processes balance load across the remaining active server computers.
1. A method of controlling the operation of a plurality of server computers wherein at least some of the plurality of server computers is located at a first facility and at least one of the plurality of server computers are located at a second facility by using a traffic manager, comprising the steps of:
the traffic manager reading a set of non-performance related variables, and wherein information about traffic between clients and the plurality of server computers passes through the traffic manager and the traffic is directed by the traffic manager;
the traffic manager determining if at least one of the non-performance related variables of the plurality of server computers can be improved by diverting tasks away from the plurality of server computers at the first facility to the at least one server computer located at the second facility; and
in response to determining that the at least one of the non-performance related variables of the plurality of server computers located at the first facility can be improved by diverting the tasks away from the plurality of server computers located at the first facility, diverting the tasks away from the servers at the first facility, at least temporarily shutting down servers at the first facility, and the traffic manager balancing load across the remaining active server computers at the second facility, wherein said non-performance variable is a duty cycle charge, such that at least one server computer located at the first facility is de-activated if its duty cycle falls below a predetermined threshold, wherein a duty cycle charge is a charge incurred whenever the at least one server computer is utilized for any duration during a predetermined period.
2. The method of claim 1 , wherein said non-performance variable is power consumption or temperature.
3. The method of claim 1 , wherein said non-performance related variable is the time of day, such that:
greater use is made of the at least one server computer located at the first facility during a first selected period; and
less use is made the at least one server computer located at the first facility during a second selected period.
4. The method of claim 1 , wherein said non-performance related variable is bandwidth for the at least one server computer located at the first facility, such that less use is made of the at least one server computer located at the first facility if a data volume accessed during a predetermined period exceeds a predetermined threshold.
5. A traffic management computer for controlling the operation of a plurality of server computers is located at a first facility and at least one of the plurality of server computers is located at a second facility, a network connecting the first facility, second facility and one or more traffic management computers, the traffic management computer comprising:
a processor; and
a non-transitory computer readable medium storing instructions that when executed by the processor cause the traffic management computer to:
perform a load balancing process to enhance operational performance;
receive a set of non-performance related variables, wherein the set of non-performance related variables includes operating temperature the at least one server computer located at the first facility and the at least one server computer is located at the second facility, and wherein information about traffic between clients and the plurality of server computers passes through the traffic management computer;
determine if the operating temperature of the at least one server computer located at the first facility can be reduced by diverting existing traffic away from the at least one server computer located at the first facility to the at least one server computer is located at the second facility;
diverting the existing traffic away from the at least one server computer located at the first facility without deactivating the at least one server computer located at the first facility in response to determining that the operating temperature of the at least one server computer located at the first facility can be reduced by diverting the existing traffic away from the server computer; and
balancing load across the remaining active server computer bandwidth at the first facility and the second facility, wherein said non-performance related data is a duty cycle charge for operation of the servers, such that the total number of active servers is reduced and workload passed to other operational servers in order to optimise the duty cycle charge for the cumulative operation of the remaining operational servers.
6. The apparatus of claim 5 , wherein said at least one non-performance related variable is selected from a selection of non-performance variable, comprising: energy consumption, bandwidth restrictions, equipment proximity, time of day and duty cycle.
7. The apparatus of claim 6 , wherein the at least one server computer located at the first facility is de-activated along with the rest of the plurality of server computers located at the first facility so as to place the whole group in a non-operational state.
8. The apparatus of claim 5 , wherein each server computer has a system clock operating at a first clock frequency and the frequency of said system clock is reduced to a second clock frequency, or a core in a multi-core processor is deactivated in response to said non-performance variable, wherein the server remains active when at the second clock frequency.
9. A computer program product, stored on a non-transitory computer readable medium, the product comprising instructions operable to cause a traffic manager to perform a method for load balancing of data sources connected to a network to enhance performance, the method comprising:
receiving a set of non-performance related variables associated with data sources located at a first facility, and wherein the traffic between clients and the data sources passes through the traffic manager;
determining if the non-performance related variables of the data sources located at the first facility can be reduced by diverting traffic away from a server computer located at the first facility to a data source that is located at a second facility; and
in response to determining that the non-performance related variables of the data sources located at the first facility can be reduced by diverting traffic away from the data source at the first facility, the traffic manager diverting the tasks away from the servers at the first facility, and balancing load across the remaining active server computers at the second facility, wherein said non-performance variable is duty cycle charge, wherein data sources are selected in order to maintain active use over an operational period while other data sources remain totally inactive over said period.
10. The computer program product of claim 9 , wherein said instructions configure the traffic manager to de-activate the plurality of server computers located at the first facility.
11. The computer program product of claim 9 , wherein said non-performance related variable is time of day, such that greater use is made of data sources resident at geographical locations that are experiencing off-peak conditions or cooler operating temperatures.
12. The computer program product of claim 9 , wherein said non-performance related variable is bandwidth, such that a data source is no-longer selected after a predetermined level of data output.