Hierarchical network, service and application function virtual machine partitioning across differentially sensitive data centers

• US 9,288,148 B1
• Filed: 10/30/2014
• Issued: 03/15/2016
• Est. Priority Date: 10/30/2014
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Krishnaswamy, Dilip, Kothari, Ravi, Gabale, Vijay
• Primary Examiner(s): Frech, Karl D
• Application Number:
  US14/528,557
• Time in Patent Office: 502 Days
• Field of Search: 709/217, 235/462.31, 235/462.15, 235/462.25
• US Class Current: 1/1
• CPC Class Codes: H04L 41/0893 Assignment of logical groups to network elements
• CPC Class Codes: H04L 41/0894 Policy-based network configuration management
• CPC Class Codes: H04L 41/0897 by horizontal or vertical scaling of resources, or by migrating entities, e.g. virtual resources or entities

Abstract

Methods, systems, and computer program products for hierarchical DCS-aware network, service, and application function VM partitioning are provided herein. A method includes partitioning multiple functions, within a set of virtual machines distributed across a hierarchical network of two or more data centers, into a first set of functions and a second set of functions, wherein the first set is associated with a higher performance sensitivity measure than the second set, and wherein said partitioning is based on a desired performance sensitivity measure associated with the functions and data center sensitivity measures of the data centers; executing the first set of functions in a first data center associated with a higher data center sensitivity measure than the one or more additional data centers; and executing the second set of functions in a second data center associated with a lower data center sensitivity measure than the first data center.

Claims

1. A method comprising the following steps:

partitioning multiple functions, within a set of virtual machines distributed across a hierarchical network of two or more data centers, into at least a first set of functions and a second set of functions, wherein the first set of functions is associated with a higher performance sensitivity measure than the second set of functions, and wherein said partitioning is based on (i) a desired performance sensitivity measure associated with the multiple functions and (ii) data center sensitivity measures provided by the two or more data centers;

executing the first set of functions in one or more of the virtual machines in a first of the two or more data centers, wherein the first data center is associated with a higher data center sensitivity measure than the one or more additional data centers in the hierarchical network of data centers; and

executing the second set of functions in one or more of the virtual machines in a second of the two or more data centers, wherein the second data center is associated with a lower data center sensitivity measure than the first data center;

wherein at least one of the steps is carried out by a computing device.

2. The method of claim 1, wherein said partitioning is carried out by a hierarchical function manager in a distributed manner across the two or more data centers.

3. The method of claim 2, wherein the hierarchical function manager is assisted by a local hierarchical function manager that manages one or more functions for a local data center among the two or more data centers.

4. The method of claim 2, wherein the hierarchical function manager triggers a request for resource allocation for one or more of the multiple functions from one data center to another data center in the hierarchical network based on the availability of resources, a cost of available resources, the performance sensitivity measure associated with the one or more functions for which resource allocations are required, and data center sensitivity measures.

5. The method of claim 1, wherein the location of the first data center is closer to a user device than the second data center.

6. The method of claim 1, wherein the data center sensitivity measure is based on one or more of a latency sensitivity measure, a bandwidth availability measure, a network availability measure, a network utilization measure, a cost of service measure, a computing resource availability measure, a storage resource availability measure, dynamic link conditions on one or more networks associated with a user or set of users, and an energy availability measure.

7. The method of claim 1, wherein the performance sensitivity measure is based on one or more of an expected round trip delay, an overall latency of response, an expected bandwidth, the number of round trips required to accomplish a task, one or more computation requirements for function execution, one or more storage requirements for function execution, one or more dynamic wireless link conditions for a mobile device that utilizes a given function, and a given cost for function execution.

8. The method of claim 1, comprising:

replicating the multiple functions across the hierarchical network of two or more data centers based on an ability of the two or more data centers to meet performance sensitivity measures associated with the multiple functions.

9. The method of claim 1, wherein one or more of the multiple functions are collapsed into one of the two or more data centers to provide one or more of (i) a reduction in latencies associated with inter-function communications, (ii) an improvement in end-to-end application performance, and (iii) differentiated services for one or more users.

10. The method of claim 1, wherein the multiple functions comprise multiple network functions associated with providing a network service in the hierarchy of two or more data centers.

11. The method of claim 1, comprising:

partitioning resources across the two or more data centers to support two or more network operators; and

providing an inter-operator tunneling service between the operator networks in the hierarchical network of two or more data centers, wherein said providing comprises establishing an encrypted data tunnel between a packet gateway virtual machine of a first of the two or more network operators and a packet gateway virtual machine of a second of the two or more network operators.

12. The method of claim 1, comprising:

classifying each of the two or more data centers based on ability to respond to one or more user devices in a given geographic area and/or a networked region.

13. The method of claim 1, comprising:

splitting one of the multiple network functions into multiple sub-functions;

partitioning the multiple sub-functions into at least a first set of sub-functions and a second set of sub-functions, wherein the first set of sub-functions is associated with a higher performance sensitivity measure than the second set of sub-functions;

executing the first set of sub-functions in one or more virtual machines in the first data center; and

executing the second set of sub-functions in one or more virtual machines in the second data center.

14. The method of claim 1, wherein (i) said partitioning, (ii) said executing the first set of network functions, and (iii) said executing the second set of network functions is performed dynamically for each of multiple users.

15. The method of claim 1, wherein (i) said partitioning, (ii) said executing the first set of network functions, and (iii) said executing the second set of network functions is performed dynamically for a subset of users from multiple users, wherein the subset of users corresponds to a given criterion, and wherein the given criterion comprises one or more quality requirements of the users, one or more operator network constraints, and/or a cost of services associated with the users.

16. The method of claim 1, comprising:

dynamically placing the set of virtual machines across the hierarchical network of two or more data centers based on one or more attributes, wherein said dynamically placing comprises replicating one more of the virtual machines across the two or more data centers.

17. The method of claim 16, wherein the one or more attributes comprise (i) availability of computing resources in each of the two or more data centers, (ii) latency requirements associated with the multiple functions, (iii) capabilities of the two or more data centers to meet one or more latency requirements, (iv) a cost of service in the two or more data centers, (v) availability of storage resources in each of the two or more data centers, (vi) availability of energy resources in each of the two or more data centers, (vii) one or more link conditions associated with one or more users being served by the two or more data centers, (viii) availability of one or more networks connected to each of the two or more data centers, and/or (ix) a load on the hierarchical network of two or more data centers.

18. The method of claim 1, comprising:

interacting with one or more physical network nodes in the hierarchical network of two or more data centers to obtain a policy for a virtual machine partitioning scheme;

decoupling enforcement of the policy to implement the policy at one or more local nodes of the hierarchical network of two or more data centers; and

enforcing the policy based on one or more context attributes, wherein the one or more context attributes comprise (i) time of day, (ii) congestion in the hierarchical network of two or more data centers, (iii) cost of service in the two or more data centers, and/or (iv) requirements of one or more users and/or applications.

19. The method of claim 1, comprising:

predicting an expected usage of resources across the two or more data centers, wherein the resources comprise one or more of a computing resource, a storage resource, a networking resource, and an energy resource, and wherein said predicting is based on one or more of a static policy input, a dynamic policy input, dynamic information of resource requirements in the two or more data centers, overprovisioning of resources based on anticipated demand, and an internal estimate of dynamic usage requirements; and

determining a future allocation of resources across the two or more data centers based on said predicting.

20. The method of claim 1, comprising:

assigning a given one of the multiple functions to a given one of the two or more data centers to support a set of users based on the data center sensitivity measure corresponding to the given data center and the performance sensitivity measure corresponding to the given function.

21. The method of claim 1, comprising:

enabling a given one of the multiple functions to execute in a software container in a given virtual machine from the set of virtual machines; and

migrating the given function between two of the two or more data centers based on (i) the data center sensitivity measures of the two data centers, (ii) the performance sensitivity measure associated with the given function, and (iii) a cost of migration associated with the given function.

22. The method of claim 1, wherein said partitioning comprises:

placing one or more cloud application virtual machine functions for a given user requiring higher performance sensitivity in a cloud data center with a higher data center sensitivity measure relative to the given user.

23. The method of claim 1, comprising:

implementing function replication with concurrent utilization of a given function at two or more data center locations in the hierarchical network of data centers.

24. The method of claim 1, wherein the set of virtual machines comprises a machine-to-machine service layer service function virtual machine and/or an application proxy function virtual machine.

25. A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to:

partition multiple functions, within a set of virtual machines distributed across a hierarchical network of two or more data centers, into at least a first set of functions and a second set of functions, wherein the first set of functions is associated with a higher performance sensitivity measure than the second set of functions, and wherein said partitioning is based on (i) a desired performance sensitivity measure associated with the multiple functions and (ii) data center sensitivity measures provided by the two or more data centers;

execute the first set of functions in one or more of the virtual machines in a first of the two or more data centers, wherein the first data center is associated with a higher data center sensitivity measure than the one or more additional data centers in the hierarchical network of data centers; and

execute the second set of functions in one or more of the virtual machines in a second of the two or more data centers, wherein the second data center is associated with a lower data center sensitivity measure than the first data center.

26. A system comprising:

a memory; and

at least one processor coupled to the memory and configured for;

partitioning multiple functions, within a set of virtual machines distributed across a hierarchical network of two or more data centers, into at least a first set of functions and a second set of functions, wherein the first set of functions is associated with a higher performance sensitivity measure than the second set of functions, and wherein said partitioning is based on (i) a desired performance sensitivity measure associated with the multiple functions and (ii) data center sensitivity measures provided by the two or more data centers;

executing the first set of functions in one or more of the virtual machines in a first of the two or more data centers, wherein the first data center is associated with a higher data center sensitivity measure than the one or more additional data centers in the hierarchical network of data centers; and

executing the second set of functions in one or more of the virtual machines in a second of the two or more data centers, wherein the second data center is associated with a lower data center sensitivity measure than the first data center.

27. A method comprising the following steps:

partitioning multiple functions, within a set of virtual machines distributed across a hierarchical network of two or more data centers and in connection with a set of multiple users, into at least a first set of functions and a second set of functions, wherein the first set of functions corresponds to a subset of one or more users associated with a given level of performance sensitivity, and wherein said partitioning is based on (i) a desired performance sensitivity measure associated with the multiple functions and (ii) data center sensitivity measures provided by the two or more data centers; and

deploying differentiated services among the set of multiple users by;

executing the first set of functions in one or more of the virtual machines in a first of the two or more data centers, wherein the first data center is associated with a higher data center sensitivity measure than the one or more additional data centers in the hierarchical network of data centers; and

executing the second set of functions in one or more of the virtual machines in a second of the two or more data centers, wherein the second data center is associated with a lower data center sensitivity measure than the first data center;

wherein at least one of the steps is carried out by a computing device.