EP2452249A1

EP2452249A1 - Method for monitoring operation of a computation network

Info

Publication number: EP2452249A1
Application number: EP10725596A
Authority: EP
Inventors: Matthias RÖBKE; Helmut Krack-Haller
Original assignee: Deutsche Telekom AG
Current assignee: Deutsche Telekom AG
Priority date: 2009-07-08
Filing date: 2010-04-23
Publication date: 2012-05-16
Also published as: DE102009032133A1; WO2011003375A1

Abstract

The invention relates to an arrangement (32) for monitoring operation of a computation network (28), which comprises a number of regionally distributed computation centres (12, 16, 20, 24), wherein it is provided that electrical power for operation of the computation centres (12, 16, 20, 24) is provided by respective regionally distributed power suppliers (14, 18, 22, 26) in a power network (30), wherein the arrangement (32, 38) has at least one power monitoring entity (34, 40, 42) which is designed to determine a potential and/or current power demand of at least one computation centre (12, 16, 20, 24), to compare those with a correspondingly regionally provided available power, and to distribute the computation power to be produced by the computation network (28) to the at least one computation centre (12, 16, 20, 24) corresponding to the available regional power.

Description

Method for controlling an operation of a

Computing network

The present invention relates to an arrangement and a method for controlling an operation of a

Computing network. Furthermore, the present invention relates to an arrangement for controlling an operation of a computer network, a computer program product and a corresponding computer program.

Background of the invention

When using renewable energies such as wind power, solar energy, hydropower, etc., the availability of which may depend on the time of the year and / or the time of day, as well as the weather, it can be regional and

time to discrepancies between available and

needed energy come. Thus, a situation is conceivable, according to which z. B. in a region A due to possibly temporally limited increased solar radiation or strong wind an oversupply of energy, but this energy is not needed in the vicinity of said region A or can be consumed.

The telecoms and IT industry has in recent years built up large computer networks or computer networks, which are generally distributed regionally. The energy demand of individual data centers within the

Computing network varies depending on the degree of utilization of the data center and is usually different in time as well as regional.

A user who accesses the computing capacity of the described computer networks usually has no knowledge of the topology of the actual computer network. Thus, it is not clear to the user whether the computing power of a data center in the region A or a

Data Center in Region B is provided. Under this aspect is an architecture of the computer network

virtualized.

Power plants of energy suppliers, distributed in different regions, and u. a. renewable energies using various new technologies

provide, produce the energy possibly regionally and temporally differently distributed, so that regionally a surplus or a supply of energy compared to the energy actually required can arise at any given time.

A control of the load distribution within the

Computer network interconnected data centers so far only in case of error taking advantage of existing redundancy mechanisms. This can mean that in the event of a partial or complete failure of a data center, at least one second data center takes over the tasks of the first data center. Summary of the invention The invention relates to an arrangement for controlling an operation of a computing network comprising a number of regionally distributed data centers. It is

provided that electrical energy to operate the data centers of each regionally distributed

Energy suppliers or energy producers of a

Energy network is provided. The arrangement has at least one energy control entity or unit. The at least one energy control entity is to

trained, a potential and / or current

Energy needs of at least one data center to be determined and with a corresponding regional existing

Energy supply to compare. In addition, the at least one energy control entity is designed to provide a computing power to be provided by the computing network in accordance with the regional energy supply or a regional energy supply

Availability of typically distribute electrical energy to the at least one data center. A distribution of the computing power or computing capacity is u. a. to carry out by carrying out the

Computing applications or computational applications and thus of computational network tasks at least one particular

Computer center is selected specifically and the

Computing applications are executed by this at least one data center. The at least one data center located in a region draws its energy

typically from at least one energy supplier from the same region. Accordingly, a region - dependent selection of the at least one energy supplier and the energy provided by the latter takes place by selecting the at least one data center regionally assigned to the at least one energy supplier.

The envisaged energy control entity can be arranged centrally. In an embodiment, the at least one energy control entity is designed, depending on the regional Ξnergieangebot between the

Data centers have a corresponding distribution of

Perform computing power, d. H. to initiate and control.

In addition, the energy control entity can be designed to determine an actual energy requirement of the respective data centers at predeterminable time intervals, and, if necessary, in coordination with the regionally available energy supply, a corresponding temporal and regional, uninterrupted transfer of the energy

To control computing power between the data centers. In a variant, the arrangement may have at least one control entity for the computing network connected or cooperating with the at least one energy control entity, which is designed to store data or data

Information about the energy requirement and / or the respective required computing power or computing capacity of the

Provide data centers and distribute computing power to data centers. This at least one

Control instance, which may be implemented as part of a management system, may be of the at least one

Energy control instance to be initialized. In addition, the arrangement can at least one with the

have at least one energy control entity connected or cooperating energy network management system, which is designed to data or

Information with parameters about the regional and / or temporal availability of energy or about the

to provide the respective energy offers. The data or information may also include parameters relating to an energy price and / or energy sources used. The data provided in this way are provided by the

Energy control instance used in a distribution of the computing power, taking into account the energy. It is also conceivable that the energy control body to plan the distribution of computing power directly with the energy suppliers and the data centers under

typically mutual exchange of data

cooperates, so that the use of tax authorities and / or management systems is not essential.

There may also be provided a plurality of decentralized energy control entities. Consequently, the arrangement can have at least one energy control instance assigned to the computing network and / or at least one energy control entity assigned to the energy network.

The invention further relates to a method for

Controlling an operation of a computing network comprising a number of regionally distributed data centers. It It is envisaged that electrical energy to operate the data centers of each regionally distributed

Energy supplier of an energy network. To carry out the method is a

Energy requirement of at least one data center determined and with a regionally available energy supply of

compared to at least one energy supplier. Furthermore, one to be provided by the computer network

Computing power according to the regional offered amount of energy or the regional available

Energy supply distributed to the at least one data center regionally.

In one embodiment of the method, energy suppliers or energy suppliers are selected as needed, with energy suppliers typically being able to be preferred which supply the energy cheaply and / or conventionally

offer renewable energy sources. In the context of the procedure, a distribution of

The respective computing capacities or computer services of the computer centers are adapted to the particular regional energy supply, which can be dependent on an energy requirement and / or consumption.

Alternatively or additionally, a distribution of

Computing services in accordance with a respective regional energy supply, a temporal availability of energy and / or depending on an energy price between the data centers are distributed in a timely manner. Besides, it is possible the computing power depending on which energy sources the energy is provided from.

The data centers are usually linked together within the computer network. The same can also be provided for the energy suppliers or their power plants within the energy network.

The data centers and energy suppliers are distributed regionally or geographically within an area that may include at least one country and / or at least part of a country. Within a region can

at least one data center and at least one

Be arranged energy supplier. Typically, a data center receives its required electrical energy from an energy supplier from its environment, typically an energy supplier, to the data center

has the smallest distance. With the invention u. a. Provided mechanisms that allow the energy consumption of

Computing networks with their geographically distributed

Intelligent data center management and thus regionally efficient control. In an embodiment, a shift of the required computer capacity for

efficient use of the locally provided

Energy offer.

The arrangement described is designed to carry out all the steps of the presented method. In doing so, individual steps of this method can also be carried out by individual components of the arrangement, usually the at least one energy control entity, are performed. Furthermore, functions of the arrangement or functions of individual components of the arrangement can be implemented as steps of the method. In addition, it is possible that steps of the method as functions of individual components of the device or the whole

Arrangement can be realized. The invention also relates to a computer program with

Program code means by means of which all steps of a presented method can be carried out when the computer program is run on a computer or a computer

corresponding arithmetic unit, in particular using a described arrangement, is executed.

A computer program product according to the invention has a computer-readable medium and a computer program stored on the computer-readable medium

Program code means adapted to carry out all steps of a described method when the computer program is executed on a computer or a corresponding arithmetic unit, in particular using a proposed arrangement.

The invention includes, in an embodiment, a method and an arrangement or instance for controlling the energy balance and shifting computer power of a number of first data centers into a number of second, others

Data centers. Thus, it is possible to have an energetic

Shift capacity more efficiently and thus shape it. Depending on the energy offered

Computer applications or computing applications that

usually performed by arithmetic operations, moved between the data centers, so that it is possible to distribute the computing power by distributing the applications or adapt.

In the context of the invention, it is provided that a regional load distribution of the electrical energy does not occur

Requirement of at least one operator of the

Energy network and thus energy suppliers, but on the basis of specifications of the customer takes place. The customer can adapt himself to regional and temporal differences in energy supply, for example with regard to electricity tariffs, by regional distribution of the computing power.

Thus, the customer, i. H. an operator of

Computing network, which itself determine the decrease of the electrical energy by the energy suppliers by concentrating the computing capacities on at least one data center, which is supplied within a region by its preferred energy supplier with electrical energy. This is done u. a. in that by the computing applications to be executed by the

Computing network are at least one

Data center serviced by the energy supplier. A selection of an energy supplier is therefore not external but internal, within the computer network by appropriate assignment of

Computing applications to at least one data center. A regional overarching component of the arrangement or authority for controlling the energy balance in

Computing networks, typically called

Energy control instance is called, is to

trained to control the temporal and regional distribution of the required computing power in coordination with each available on-site energy and thus to control and / or regulate. The energy control entity determines this permanently or in predeterminable

Time intervals the actual energy needs of at least one data center, which is part of the distributed

Computing network is. The energy control body is trained to compare this energy demand with the locally available local energy supply.

In the case of a possible quantitative subset of required energy for a first local data center A, the energy control entity checks which second data center B located in another region can take up additional computing power and shifts the operation of the regionally comprehensive and uninterrupted power supply

respective applications from the data center A

Capacity-appropriate in the determined data center B. This shift in the determined data center B is carried out by a regional overarching control authority for the computing and / or energy network, of the

Energy control instance is initialized. One

Management system for computing networks is, for example, in the

Error case used to implement redundancy mechanisms between the regionally distributed data centers. The same applies in the case of a quantitative oversupply of energy in the region of a certain data center in one

Embodiment of the invention. If such an oversupply of energy exists, then with the help of the energy control authority, computing services can be regionally provided by selected execution of computing applications from other regions

Data center will be relocated to the quantitative

To compensate for excess supply of energy. The energy control body usually has knowledge about the energy balance of a regionally distributed computer network as well as the actual available energy in a certain region. To get data about the available energy and the required computing capacity or performance is the

Energy Control Instance with the least one

Management system of at least one energy supplier or - suppliers and connected to the at least one control entity of the computer network. Is from the

Energy control entity an imbalance between

actually needed and regionally available energy or any other mismatch in terms of energy actually needed and a regionally available energy

Energy supply, for example, in terms of an applicable

Energy tariffs, found so causes the

Energy control entity according to a procedure, such as in redundancy mechanisms for the distribution of

Computing services may be provided, a suitable interruption-free shift of the computing power. Triggering factors for such a shift of

Computing power can be provided in addition to the regional quantitative oversupply of energy, unbalanced power and / or computing networks; different energy tariffs in different regions and so on.

By using the geographically comprehensive

Energy control entity _, the distributed data center operator can, for reasons of network security,

possibly swivel from data center A to data center B, if necessary also manually.

Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings.

It is understood that the features mentioned above and those yet to be explained below can be used not only in the particular combination indicated, but also in other combinations or in isolation, without departing from the scope of the present invention.

Brief description of the drawings

The invention is based on embodiments in the

Drawings shown schematically and will be described in detail below with reference to the drawings

described.

Figure 1 shows a schematic representation of an area with a first embodiment of an inventive

Arrangement for implementing a first embodiment of a inventive method for a distributed

Computing network.

Figure 2 shows a schematic representation of the presented in Figure 1 area with a second embodiment of an inventive arrangement.

Detailed description of the drawing The figures become coherent and comprehensive

described. The same reference numerals designate the same

Components.

FIG. 1 shows a schematic representation of an example of a region 2 which comprises a first geographical region 4, a second geographical region 6, a third geographic region 8 and a fourth geographic region 10

having. Within the area 2, the first region 4 is assigned a first data center 12 and a first energy supplier 14. The second region 6 is a second one

Data center 16 and a second energy supplier 18 assigned. Accordingly, the third region 8, a third data center 20 and a third energy supplier 22 and the fourth region 10, a fourth data center 24 and a fourth energy supplier 26 are assigned.

It is envisaged that all of these data centers 12, 16, 20, 24 will be linked across regions

Computing network 28 are joined together. The

Energy suppliers 14, 18, 22, 26 are combined to form a common energy network 30. The data centers 12, 16, 20, 24 are formed within the computing network 28 'for performing computing applications. In this case, computing applications and thus computing power depending on the load within the

Computing network 28 between the data centers 12, 16, 20, 24 distributed. Within the energy network 30 energy from the 14nergielieferanten 14, 18, 22, 26

produced to the network 28 to operate the

Data centers 12, 16, 20, 24 is provided.

A first shown schematically in Figure 1

Embodiment of an inventive arrangement 32 comprises a first embodiment of a central

Energy control entity 34. This energy control entity 34 is designed to be controlled by a control entity 36 of the

Computing network 28 with the computing network 28

co. Between the energy control entity 34 and the control entity 36 of the computing network 28, information about a regional utilization of the

Data centers 12, 16, 20, 24 exchanged. Furthermore, the energy control entity 34 is configured to communicate with the

Energy Network 30 and interact from that

Energy Network 30 here information about an electric offered by the respective energy suppliers

To get energy.

In FIG. 1, the energy control entity 34 is formed as a link between the energy network 30 of the energy suppliers 14, 18, 22, 26 and the operators of the data centers 12, 16, 20, 24 of the computing network 28. The

Energy Control Instance 34 is here in relation to the Energy network 30 and the computing network 28 is a central unit.

The connection of the energy control entity 34 to a

Control system of the energy suppliers 14, 18, 22, 26 as well as those of the operator of the data centers 12, 16, 20, 24 within the computing network 28 takes place in the

present embodiment via known,

industry-standard mechanisms and the associated communication protocols.

In Figure 2, the already schematically illustrated with reference to Figure 1 area 2 is shown again. In addition, Figure 2 shows a schematic representation of a second

Embodiment of an inventive arrangement 38, which differs from the arrangement 32 shown in FIG

different. The second embodiment of the

Arrangement 38 according to the invention comprises a first

Power control instance 40, which is associated with the computing network 28, and a second energy control entity 42, which is associated with the energy network 30. A

Embodiment of the method according to the invention is achieved by co-operation of the two

Energy control bodies 40, 42 with exchange of

Information between the two energy control instances 40, 42 executed.

As an alternative to a central energy control entity 34, as shown in FIG. 2, in a further refinement, two separate, distributed energy control entities 40, 42 may also be provided. In this case, a first monitors Energy Control Authority 42 the energy balance in the energy network 30 and a second, others

Energy Control Panel 40 the resource balance in the

Computing Network 28.

If, in this embodiment, there is a quantitative oversupply or supply shortage in the energy network 30 of the energy suppliers 14, 18, 22, 26, this event, together with a query, reaches which data centers 12, 16, 20, 24 in which region 4, 6, 8 , 10 the computing power is to be switched or diverted to the

Energy Control Instance 40 in the computing network 28

signaled. The energy control instance 40 in

Computing network 28 evaluates the request and initiates the requested switchover or signals the

Power Control Instance 42 in Energy Network 30 of the

Energy suppliers 14, 18, 22, 26, that switching with the supplied parameters to a selected region 4, 6, 8, 10 is not possible. The supervisory authority in the

Power grid or energy network 30 can then possibly the

Repeat request with a modified target region. The described functions of the two energy control units 40, 42 can also typically only one

Energy control instance 34, as shown in Figure 1, are performed.

The at least one energy control entity 34, 40, 42 can be described in FIGS. 1 and 2

Embodiments use further data for the decision of switching or relocation. So can

statistical values, such as For example, a predicted one Power consumption of a region 4, 6, 8, 10 or an expected computing power of a data center 12, 16, 20, 24, are used. Furthermore, it is possible to define suitable threshold values for the switching between data centers 12, 16, 20, 24, in order, for example, to avoid so-called hysteresis effects.

In both presented with reference to Figures 1 and 2

Embodiments, a selection of at least one energy supplier 14, 18, 22, 26 and thus the energy provided by him by selecting at least one data center 12, 16, 20, 24, the regional of at least one energy supplier 14, 18, 22, 26 with energy

is supplied. A selection of at least one

Data Center 12, 16, 20, 24 is by assignment of

Computing applications or computing applications and thus of computing power to the at least one data center 12, 16, 20, 24 made. The at least one selected data center 12, 16, 20, 24 obtains its required

Energy region dependent on at least one

Energieiieferanten 14, 18, 22, 26, which typically in the same region 4, 6, 8, 10 as the at least one

Data Center 12, 16, 20, 24 is located.

Claims

claims

1. Arrangement for controlling an operation of a

Computing Network (28), a number regionally

distributed data centers (12, 16, 20, 24), wherein it is provided that electrical energy for operating the data centers (12, 16, 20, 24) of respective regionally distributed energy suppliers (14, 18, 22, 26) of an energy network ( 30) is provided,

wherein the assembly (32, 38) at least one

Energy control instance (34, 40, 42), which is adapted to determine a potential and / or current energy needs of at least one data center (12, 16, 20, 24) to compare with a corresponding regionally available energy supply and one of the computer network (28) to be provided computing power according to the regional energy supply to the at least one data center (12, 16, 20, 24) to distribute.

2. Arrangement according to claim 1, wherein the at least one energy control entity (34, 40, 42) is adapted to, depending on each regionally present

Energy supply a corresponding distribution of

To initiate and control computing power between the data centers (12, 16, 20, 24).

3. Arrangement according to claim 1 or 2, wherein the

at least one energy control entity (34, 40, 42) is designed to each of an actual energy requirement of the respective at predetermined intervals

Data centers (12, 16, 20, 24) to determine, and in accordance with the respectively existing regional

Energy supply, if necessary, a corresponding temporal and regional, uninterrupted relocation of the

Computing power between the data centers (12, 16, 20, 24) to regulate.

4. Arrangement according to claim 1, 2 or 3, which has at least one with the at least one energy control entity (34, 40, 42) cooperating control entity (36) for the computing network (28), which is adapted to data on a respectively required computing capacity the data centers (12, 16, 20, 24).

5. Arrangement according to one of the preceding claims, the at least one with the at least one

Energy Control Authority (34, 40, 42) Cooperating Energy Network Management System (30)

, which is designed to provide data on the respective regional energy supply

provide.

6. Arrangement according to one of the preceding claims, the at least one the computing network (28) associated energy control entity (34, 40, 42) and at least one energy network (30) associated

Energy control instance (34, 40, 42).

7. Method for controlling an operation of a

Computing Network (28), a number regionally

wherein for carrying out the method a

Energy requirement of at least one data center (12, 16, 20, 24) is determined and compared with a regionally available energy supply of at least one energy supplier (14, 18, 22, 26), and in which one of the computing network (28) to be provided computing power accordingly the regionally available energy supply to the at least one data center (12, 16, 20, 24) is distributed.

8. The method of claim 7, wherein a distribution of the respective computing power to the data centers (12, 16, 20, 24) is adapted to the respective regional energy supply.

9. The method according to any one of claims 7 or 8, wherein the respective computing power corresponding to a respective regionally available energy supply between the data centers (12, 16, 20, 24) in time

be distributed or relocated in a controlled manner.

10. Computer program with program code means, by means of which all steps of a method according to one of claims 7 to 9 can be performed when the computer program on a computer or a

corresponding arithmetic unit, in particular in an arrangement (32, 38) according to one of claims 1 to 6, is executed.

11. Computer program product with a computer readable

Medium and a computer program stored on the computer-readable medium

Program code means adapted to carry out all the steps of a method according to any one of claims 7 to 9, when the

Computer program on a computer or a computer