Energy management for signal processing means TECHNICAL FIELD

The present invention relates to a node in a communication system. The node comprises at least one signal processing means and is arranged to receive information regarding future signal processing actions at said signal processing means. Said information is received a certain time before an action time when said actions are taken.

The present invention also relates to method for energy handling at a node in a communication system. The method comprises the step of receiving information regarding future signal processing actions at signal processing means in the node a certain time before an action time when said actions are taken.

BACKGROUND

Today, the need for saving energy has increased, both for economical and environmental reasons. There is therefore a desire to save energy in many fields of technology, but at the same time there is a desire to maintain an expected service quality. Many types of equipment, such as radio base station (RBS) equipment, comprise signal processing devices. Software signal processing applications are not often exclusively static environments; the dynamics of a radio base station baseband application is particularly multidimensional and constantly shifting. The need for processing resources can vary greatly, depending on the current user scheduling - valid for one millisecond, before it instantly changes again - while the requirements on processing latency are continuously tight and rigorous. In systems adapted for saving energy, as many as feasible of the processing devices shall be turned off, or put to a sleep mode. For an RBS, the power management of today can be divided into three typical cases:

1 ) All resources are always available. In this case there is a high energy consumption, but no unnecessary lateness.

2) Offline optimization. In this case, the best power utilization is determined offline for a corresponding predetermined load case. This means that a number of load cases are predetermined before the RBS is set on line, each load case being associated with a certain power utilization mode. When the

RBS is running, a certain load case among the predetermined load cases is determined to exist at a certain moment, and then the corresponding power utilization mode is applied. At another moment, the load may be of such a kind that another load case is determined to exist, leading to that another power utilization mode is applied. In this way, the most suitable power utilization mode case is chosen continuously as the load case of the RBS changes. However, it is difficult, or even impossible to foresee all possible load cases.

3) Online optimization. In this case, the best power utilization is continuously determined when the RBS already is running and immediately applied to the real time system. However, as the load situation is dynamically changing, this technique will always lead to either increased lateness, or unnecessary energy waste. It is therefore a desire to provide a device and method for power management technology at a signal processing device that offers a possibility to optimally minimize energy consumption and latency at the same time.

SUMMARY

It is an object of the present invention to provide a device and method for power management technology at a signal processing device that offers a possibility to optimally minimize energy consumption and latency at the same time. Said object is obtained by means of a node in a communication system. The node comprises at least one signal processing means and is arranged to receive information regarding future signal processing actions at said signal processing means. Said information is received a certain time before an action time when said actions are taken. The node is arranged to adjust the degree of operation for said signal processing means in dependence of said information.

Said object is also obtained by means of a method for energy handling at a node in a communication system. The method comprises the steps of:

- Receiving information regarding future signal processing actions at signal processing means in the node a certain time before an action time when said actions are taken.

- Using said information for adjusting the degree of operation for said signal processing means.

According to an example, said information comprises scheduling grants for user terminals that are served by the base station.

According to another example, the node is arranged introduce at least one auxiliary task, corresponding to at least one resource, to said signal processing means a certain time before the action time when said actions are to be taken. This enables said signal processing means to have a certain desired degree of operation at the action time when said actions are to be taken. Alternatively, the node is arranged to decrease the time required to at least partly turn on said signal processing means before said actions are to be taken.

More examples are disclosed in the dependent claims. A number of advantages are obtained by means of the present invention; mainly signal processing means will be able to respond to critical tasks with the lowest possible latency, while at the same time minimizing energy waste. The signal processing means is enabled to optimize its behavior in any desired way, compromising between energy consumption and task responsiveness.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will now be described more in detail with reference to the appended drawings, where:

Figure 1 shows a schematic view of a node in a wireless communication system; Figure 2 shows a diagram of resources versus time according to a first example;

Figure 3 shows a diagram of resources versus time according to a second example; and Figure 4 shows a flowchart for a method according to the present invention.

DETAILED DESCRIPTION

With reference to Figure 1 , showing a first example, there is a node 1 , here in the form of a radio base station (RBS) in a wireless communication system 2, for example an LTE (Long Term Evolution) system, a GSM (Global System for Mobile communications) system or a WCDMA (Wideband Code Division Multiple Access) system. The RBS 1 comprises an air interface scheduling device 7 that is arranged for scheduling the air interface 15 for user terminals 4, 5, 6 of different types such as cell phones and laptops. The RBS 1 comprises a plurality of signal processing means 3, where the air interface scheduling device 7 is connected to the signal processing means 3 and is arranged to determine which user terminal 4, 5, 6 that will be granted use of what air interface resources at what time.

The RBS further comprises a task scheduling device 14 that is connected to the air interface scheduling device 7 and the signal processing means 3, the task scheduling device 14 being arranged to control the signal processing means 3 by providing resource scheduling instructions. The RBS 1 is arranged to transmit the scheduling grants provided from the air interface scheduling device 7 to the user terminals 4, 5, 6 concerned. This information will then be used by each user terminal 4, 5, 6 for transmission and reception.

With reference also to Figure 2, showing a first bar diagram of resources versus time, where the amount of processing resources 3a, 3b that are used of the signal processing means 3 at a certain time is indicated with a corresponding bar, this means that the RBS 1 is arranged to receive information regarding future signal processing actions A at said signal processing means 3. The signal processing actions A correspond to the processing resources indicated with bars within the indicated span A. The received information in question is received a certain time before said actions A are to be taken. According to the present invention, the RBS 1 is arranged to adjust the degree of operation for said signal processing means 3 in dependence of said information. This is made possible by having air interface scheduling information in the form of at least part of the scheduling grants transferred to the task scheduling device 14, which in this way acquires foreknowledge of the work to be done, the future load, typically a few milliseconds in advance. The task scheduling device 14 may then adjust the degree of operation, for example by turning on or turning off, the signal processing means 3; 3a, 3b.

The present invention is thus based on use of the received information, constituting foreknowledge in order to make good online decisions on what recourses to turn on or off in the RBS. As much a priori information as possible is used to predict the future need for processing capability. With this prediction, the conflicting demands for latency and energy consumption can be optimized. The amount of energy that could be saved in this way is very substantial.

More in detail, the air interface scheduling device 7 is arranged to perform scheduling of the radio interface, i.e. to decide whom to transmit to on an up-link, and in what way. The scheduling information is then sent to the user terminals 4, 5, 6 in question via a down-link, and at least partly also sent to the task scheduling device 14. Thus the RBS 1 will comprise advance knowledge regarding in what signal processing work that has to be done in an up-link receiver 3a and downlink processing equipment 3b, the up-link receiver 3a and downlink processing equipment 3b being comprised in the signal processing means 3. The downlink processing equipment 3b also comprises information about input user data buffer sizes that can be used to predict the processing need.

For the task scheduling device 14, it is possible to use a combination of number of active users, buffer status and classes of user terminal to estimate the upcoming workload for both uplink and downlink.

The above information is used to make the RBS 1 adjust its available resources in the form of said signal processing means 3 to minimize energy consumption and latency. The task scheduling device 14 will be able to switch on signal processing means 3 in advance and even start to perform task scheduling in advance, and thus achieve both minimum latency and minimum energy consumption at the same time.

With reference to Figure 2, a first example will now be described. In this example, the a priori information according to the above is provided to the signal processing means 3 in an indirect manner. This is for example the case if the task scheduling device 14 has predetermined settings regarding wake-up time for processing resources 3a, 3b that are not possible to alter dynamically. The task scheduling device 14 is arranged to create a plurality of auxiliary tasks, zero-time tasks, corresponding to a desired certain increase 8, 9, 10, 1 1 of the use of resources at the signal processing means 3, due to the knowledge of an upcoming processing resource need at a certain action time T. These auxiliary tasks are provided to the signal processing devices 3. The auxiliary tasks are first introduced to the signal processing devices 3 at a certain time t _L +tR before the action time T, compensating for the system's wake-up inertia that here is represented by the individual processing resources' wake-up latency t _L and their combined wake-up rate t _R , respectively. This procedure would cause a number of signal processing resources 3a, 3b at the signal processing means 3 to have exited a precious low-energy state at the action time T when they are needed, providing zero-latency handling of the application's tasks upon arrival at the action time T. The number N may indicate the desired number of signal processing resources 3a, 3b that are ready for action at the action time T.

Generally, the node 1 is arranged to introduce at least one auxiliary task to the signal processing means 3 a certain time t _L +tR before the action time T when said actions A are to be taken, enabling said signal processing means 3 to have a certain desired degree of operation at the time T when said actions A are to be taken.

This method could be implemented as a totally independent add-on to a previously working system, not affecting the design of the existing resource scheduler.

With reference to Figure 3, a second example will now be described. Figure 3 shows a second bar diagram of resources versus time, where the amount of resources that are used of the signal processing means 3 at a certain time is indicated with a corresponding bar. The signal processing actions A that are to be taken at a certain action time T correspond to the resources indicated with bars within the span A. In this example, the a priori information according to the above is provided to the signal processing means 3 in a more direct manner. This is for example the case if the task scheduling device 14 has settings regarding wake-up time for processing resources 3a, 3b that may be altered dynamically. The settings of the task scheduling device 14 are then dynamically modify, reducing a wake-up latency of the signal processing resources 3a, 3b. Energy-saving modes are then not allowed to be entered when an increased processing resource need is imminent at a certain action time T. The modification is performed a first time t _A before said action time T.

Alternatively the distributed wake-up rate could be increased by letting processing resources activate more neighboring resources from low-energy states, at the time for an anticipated increased resource need.

This means that processing means 3; 3a, 3b that normally have a certain start-up time due to a sleep mode or other low-energy state, now have this low-energy state removed or at least reduced, allowing a quick start of the needed signal resources corresponding to the actions A at the action time T.

Generally, the node 1 is arranged to decrease the time t _A required to at least partly turn on said signal processing means 3 before the actions A are to be taken, enabling said signal processing means 3 to have a certain desired degree of operation at the action time T when said actions A are to be taken.

As soon as the critical, latency-sensitive tasks have been initiated, the settings of the task scheduling device 14 are re-set to their default settings, where energy saving may be of higher priority than swift response to sudden task bursts. This re- modification is performed a second time t _B after said action time T.

A real-time resource scheduler in a multi-core environment is most often nothing but a just-in-time resource scheduler, mapping tasks to be executed onto available processing resources to execute them on. The tasks may be placed in one or more global or local task queues, and processing resources may be immediately responding when they are active; or forced to be brought active from a previous low- energy state. The way the a priori information - upon which this invention is based - can be used, depends on how the existing resource scheduler is implemented.

With reference to Figure 8, the present invention relates to a method for energy handling at a node 1 in a communication system 2. The method comprises the steps of: 12: receiving information regarding future signal processing actions A at signal processing means 3 in the node 1 a certain time before an action time T when said actions A are taken; and

13: using said information for adjusting the degree of operation for said signal processing means 3.

The present invention is not limited to the examples above, but may vary freely within the scope of the appended claims. For example, the signal processing means 3 may simply be turned on or turned off, or alternatively, a certain degree of operation of the signal processing means 3 may be set.

The number of auxiliary tasks, and the nature of the actions A to be taken, may of course vary. The air interface scheduling device 7 and/or the task scheduling device 14 may, or may not, be comprised in the signal processing means 3.

The bar diagrams in Figure 2 and Figure 3 should not be regarded as showing exact time measures or resource use, but only to be indicative of understanding the present invention.

Although the description above has related to a RBS in a wireless communications system, the present invention may relate to any type of node 1 in a communication system 2, where communication either is wireless and/or via some type of wire such as copper or fiber. The node may be a repeater device or a user terminal that is communicating with another device.

As another example, the node may be in the form of a media player of similar that via fiber cable or copper cable receives media data and also data comprising information regarding future signal processing actions A at the media player. By means of this information, the media player may adjust the degree of operation for its signal processing means 3 to the upcoming processing needs, enabling energy to be saved in a similar way as described for the RBS. In this context, signal processing means may be any type of means that processes information or at least are participating in the process of processing information such as for example coding devices, decoding devices, amplifiers, active filters, switches, transmitters, receivers, microprocessors, memories, digital circuit elements, analogue circuit elements, active microwave components etc.

Furthermore, the node may be any type of device in a communication system, where the device is arranged to process any type of signals and/or data, where the device also is arranged to receive information regarding future signal processing actions A.

The task scheduling device 14 may be arranged to provide control of the signal processing means 3 even in the absence of air interface scheduling information from the air interface scheduling device 7. Then the task scheduling device 14 will be arranged to maintain an estimated degree of operation.