# US 2013/0297770 A1 · Intelligent Data Center Cluster Selection

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2013/0297770 A1
- Document ID: US20130297770A1
- Status: Published application
- Kind: A1
- Application number: 13/462,290
- Filed: 2012-05-02
- Published: 2013-11-07
- Earliest claimed: 2012-05-02
- Assignees: Futurewei Technologies, Inc.
- Inventors: Hong Zhang
- Web page: https://patentgrep.com/patent/US20130297770A1

## CPC

- G06F 9/505: {considering the load}
- H04L 67/1008: based on parameters of servers, e.g. available memory or workload (monitoring of computer activity G06F11/30)
- H04L 67/1023: based on a hash applied to IP addresses or costs
- H04L 67/1025: Dynamic adaptation of the criteria on which the server selection is based
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A method and technique for autonomous selection of a Data Center Cluster (DCC) for fulfilling a cloud computing service request, including a technique for grouping data centers (DCs) in a cloud network according to a ranking of eligible DCCs based on selection criteria. In various embodiments, the selection criteria may include a cluster performance metric, a cluster resource equivalence metric, a balance of resource performance metric, a DCC load index, or combination thereof. Other aspects include techniques for computing/determining each of the selection criteria.

## Claims

### Claim 1

In a control server of a network configured to identify a set of candidate data center clusters (DCCs) for satisfying a service request from a user, a method for grouping data centers (DCs) in a cloud network, the method comprising:
receiving a cloud computing service request from a user;
identifying a set of candidate DC clusters (DCCs) for satisfying the cloud computing service request, wherein satisfying the cloud computing service request requires providing a plurality of resources to the user;
determining selection criteria for the set of candidate DCCs, wherein the selection criteria includes one or more of a network communication cost, a cluster performance metric, a cluster resource equivalence metric, a balance of resource performance metric, and a DCC load index; and
ranking the set of candidate DCCs based on the selection criteria, wherein a highest ranked one out of the set of candidate DCCs is selected to satisfy the cloud computing service request.

### Claim 2 (depends on 1)

The method of claim 1, wherein the selection criteria includes the network communication cost, and wherein determining the selection criteria comprises:
determining a first network communication cost for a first one of the set of candidate DCCs, the first network communication cost comprising a DC-to-user cost component associated with the first candidate DCC and a DC-to-DC cost component associated with the first candidate DCC.

### Claim 3 (depends on 2)

The method of claim 2, wherein determining the first network communication cost comprises computing the DC-to-user cost component according to the formula:
Cost
DC

-

to

-

User
=
∑
k
=
1
m

C
k
*
V
k
,
where m is equal to a number of resources in the plurality of resources, Ck is a cost for transporting a unit of data between the user and a member DC providing an instant resource (resource-k), and V k is a volume of data that must be communicated between the user and the member DC providing the resource-k while the resource-k is being provided.

### Claim 4 (depends on 2)

The method of claim 2, wherein determining the first network communication cost comprises computing the DC-to-DC cost component according to the formula:
Cost
DC

-

to

-

DC
=
∑
i
=
1
,
j
=
i
+
1
m
-
1

C
r
i

r
j
*
V
r
i

r
j
,
where m is equal to a number of resources in the plurality of resources, Crirj is a cost for transporting a unit of data between a first member DC providing a first instant resource (resource-i) and a second member DC providing a second instant resource (resource-j), and Vrirj is a volume of data that must be communicated between the first member DC providing the resource-i and the second member DC providing the resource-j when providing one or both of the resource-i and the resource-j.

### Claim 5 (depends on 1)

The method of claim 1, wherein the selection criteria includes the cluster performance metric, and wherein determining the selection criteria comprises:
determining a first cluster performance metric for a first one of the set of candidate DCCs, the cluster performance metric comprising an average performance index for all member DCs of the first candidate DCC.

### Claim 6 (depends on 5)

The method of claim 5, wherein determining a first cluster performance metric comprises determining the average performance index using one or more of a networking status and performance metric, a compute status and performance metric, and a storage status and performance metric.

### Claim 7 (depends on 1)

The method of claim 1, wherein the selection criteria includes the cluster resource equivalence metric, and wherein determining the selection criteria comprises:
determining a first cluster resource equivalence metric for a first one of the set of candidate DCCs, the cluster resource equivalence metric comprising a load variation for a member DC of the first candidate DCC, the load variation corresponding to a variation between two or more resource loads of the member DC as would result from selecting the first candidate DCC to satisfy the cloud computing service request.

### Claim 8 (depends on 7)

The method of claim 7, wherein the load variation is indicative of a degree to which one of the member DC's available resources would be depleted in relation to another one or more of the member DC's available resources as a result of selecting the first candidate DCC to fulfill the cloud computing service request.

### Claim 9 (depends on 1)

The method of claim 1, wherein determining one or more selection criteria for the set of candidate DCCs comprises determining a first balance of resource performance metric for a first one of the set of candidate DCCs, the first balance of resource performance metric being based on a performance variation between two or more of the resources provided by member DCs of the first candidate DCC and corresponding to a service latency that would result from that performance variation.

### Claim 10 (depends on 1)

The method of claim 1, wherein the selection criteria includes the DCC load index criteria, and wherein determining the selection criteria comprises:
determining a first DCC load index criteria for a first one of the set of candidate DCCs, the DCC load index criteria comprising an average load value of different types of the resources of the first candidate DCC as would result from selecting the first candidate DCC to satisfy the cloud computing service request.

### Claim 11 (depends on 1)

The method of claim 1 further comprising:
determining that none of the DCs in the cloud network are capable of individually providing the plurality of resources prior to identifying the set of DCCs.

### Claim 12 (depends on 1)

The method of claim 1 further comprising:
selecting the highest ranked one out of the set of candidate DCCs to satisfy the cloud computing service request.

### Claim 13 (depends on 1)

The method of claim 1 further comprising:
providing a selection information to the user, the selection information comprising a list of the ranked candidate DCCs and a list of the selection criteria,
wherein receiving the selection information causes the user to select the highest ranked one out of the set of candidate DCCs to satisfy the cloud computing service request.

### Claim 14 (depends on 1)

The method of claim 1, wherein each candidate DCC in the set of candidate DCCs comprises two or more of the DCs that are collectively capable of providing the plurality of resources necessary for satisfying the cloud computing service request.

### Claim 15

In a control server of a network configured to identify a set of candidate data center clusters (DCCs) for satisfying a service request from a user, a method for grouping data centers (DCs) in a cloud network, the method comprising:
receiving a cloud computing service request from a user;
identifying a plurality of resources for satisfying the cloud computing service request;
identifying a plurality of candidate DC clusters (DCCs), each of which is capable of satisfying the cloud computing service request;
computing a plurality of network communication costs each of which corresponding to a unique one of the plurality of candidate DCCs; and
selecting one of the plurality of candidate DCCs that has the lowest network communication cost, wherein the selected candidate DCC is used to satisfy the cloud computing service request.

### Claim 16 (depends on 15)

The method of claim 15, wherein the plurality of network communication costs comprise a first network communication costs corresponding to a first one of the plurality of candidate DCCS, and wherein computing the first network communication cost comprises:
computing a DC-to-user cost component associated with the first candidate DCC;
computing a DC-to-DC cost component associated with the first candidate DCC; and
adding the DC-to-user cost component and the DC-to-DC cost component to obtain the first network communication cost.

### Claim 17 (depends on 16)

The method of claim 16, wherein the first candidate DCC comprises a member DC for providing a first one of the plurality of resources, and
wherein computing the DC-to-user cost component comprises determining a network cost associated with transporting traffic between the member DC and the user as would result if the first candidate DCC was selected to fulfill the cloud computing service request, the traffic being transported as a result of providing the first resource.

### Claim 18 (depends on 16)

The method of claim 16, wherein the first candidate DCC comprises a first member DC for providing a first one of the plurality of resources and a second member DC for providing a second one of the plurality of resources, and
wherein computing the DC-to-DC cost component comprises determining a network cost associated with transporting traffic between the first member DC and the second member DC as would result if the first candidate DCC was selected to fulfill the cloud computing service request, the traffic being transported as a result of providing one or both of the first resource and the second resource.

### Claim 19 (depends on 15)

The method of claim 15 further comprising determining that none of the DCs in the cloud network are capable of individually providing each of the plurality of resources before identifying the plurality of candidate DCCs.

### Claim 20

A non-transitory computer program product for use in a control server of a network configured to identify a set of candidate data center clusters (DCCs) for satisfying a service request from a user, the computer program product configured to perform a method for grouping data centers (DCs) in a cloud network by comprising computer executable instructions stored in logical hardware, on computer readable storage media, or a combination thereof, such that when a processor associated with the control server executes said computer executable instructions it causes the control server to perform the following:
receive a cloud computing service request from a user;
identify two or more groups of DCs for satisfying the cloud service request, wherein satisfying the cloud computing service request requires providing a plurality of resources to the user;
identify selection criteria for each group of DCs, wherein the selection criteria includes one or more of a network communication cost, a cluster performance metric, a cluster resource equivalence metric, a balance of resource performance metric, and a DCC load index; and
rank each group of DCs based on the selection criteria, wherein a highest ranked one of the two or more groups of DCs is selected to satisfy the cloud computing service request.

### Claim 21 (depends on 20)

The non-transitory computer program product of claim 20, wherein the selection criteria comprises the cluster performance criteria, and wherein the cluster performance criteria is indicative of an average performance for all member DCs in the corresponding group of DCs.

### Claim 22 (depends on 20)

The non-transitory computer program product of claim 20, wherein the selection criteria comprises the cluster resource equivalence criteria, and wherein the cluster resource equivalence criteria is indicative of an average load variation for all member DCs of the corresponding group of DCs.

### Claim 23 (depends on 20)

The non-transitory computer program product of claim 20, wherein the selection criteria comprises the balance of resource performance equivalence criteria, wherein the balance of resource performance equivalence criteria is indicative of a performance variation between two or more of the resources provided by member DCs in the corresponding group of DCs.

### Claim 24 (depends on 20)

The non-transitory computer program product of claim 20, wherein each group of DCs comprises two or more member DCs that are collectively capable of providing the plurality of resources.
