Placing objects on hosts using hard and soft constraints

• US 9,075,661 B2
• Filed: 10/20/2010
• Issued: 07/07/2015
• Est. Priority Date: 10/20/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Hopmann, Alexander, Windle, Marc Keith, Simek, Patrick, Lerma, Erick Raymundo, Rosenthal, Dan
• Primary Examiner(s): HOANG, HIEU T
• Application Number:
  US12/908,681
• Publication Number:
  US 20120102199A1
• Time in Patent Office: 1,721 Days
• Field of Search: 709/224, 709/225, 709/226, 718/1
• US Class Current: 1/1
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/5016 the resource being the memory
• CPC Class Codes: G06F 9/5022 Mechanisms to release resources
• CPC Class Codes: G06F 9/5044 considering hardware capabilities
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5055 considering software capabilities, i.e. software resources associated or available to the machine

Abstract

Objects are placed on hosts using hard constraints and soft constraints. The objects to be placed on the host may be many different types of objects. For example, the objects to place may include tenants in a database, virtual machines on a physical machine, databases on a virtual machine, tenants in directory forests, tenants in farms, and the like. When determining a host for an object, a pool of hosts is filtered through a series of hard constraints. The remaining pool of hosts is further filtered through soft constraints to help in selection of a host. A host is then chosen from the remaining hosts.

Claims

1. A method for placing objects on hosts in an online service, comprising:

determining, using a processor, when to place an object on one of a plurality of hosts in the online service, wherein the online service comprises databases, virtual machines (VMs) and physical machines, wherein the physical machines are arranged in farms that perform roles in a network;

obtaining placement constraints for the object that comprise hard constraints and soft constraints, wherein application of a hard constraint eliminates one or more of the hosts from being selected on which to place the object, and wherein soft constraints determine which hosts are more suitable than other hosts;

applying the hard constraints against hosts within the network that results in a pool of remaining hosts, wherein the hard constraints include high availability constraints, wherein applying the high availability constraints eliminates one or more hosts to ensure that objects having redundant roles within the network are placed on different hosts;

applying the soft constraints against the pool of remaining hosts that results in a pool of selection hosts;

automatically selecting a host on which to place the object from the pool of selection hosts; and

automatically deploying the object on the selected host.

2. The method of claim 1, wherein the object is at least one of:

a virtual machine (VM), a tenant or a database.

3. The method of claim 1, wherein the selected host is at least one of:

a physical machine, a database, a farm, a rack that holds physical machines, or a virtual machine (VM).

4. The method of claim 2, wherein when the object is a VM and the plurality of hosts are physical machines, applying the hard constraints comprises applying two or more of the following constraints against each host:

host is located within the same network as the farm the VM is being deployed for;

host is of the correct type based on a role of the VM;

host is in the running state;

the host includes capacity to accommodate the VM;

or host does not already have a VM of the same role and farm.

5. The method of claim 1, wherein the constraints are used to prevent any one of the hosts from being overloaded with objects.

6. The method of claim 1, wherein when the object is a virtual machine and the plurality of hosts are physical machines, the hard constraints comprise applying two or more of the following constraints against each host:

host is of a correct type;

host is in the running state;

host includes capacity to accommodate the object;

or host does not already have an object of the same type.

7. The method of claim 1, wherein the soft constraints are optimizations that when applied to the pool of remaining hosts spread the deployment of objects of a same type across different hosts.

8. The method of claim 1, wherein objects of a larger size are placed before objects having a smaller size.

9. The method of claim 1, wherein the hard constraints further include performance constraints.

10. A computer-readable storage medium, excluding a signal, having computer-executable instructions for deploying objects on hosts in an online service, comprising:

determining when to place an object on one of a plurality of hosts in the online service, wherein the online service comprises databases, virtual machines (VMs) and physical machines, wherein the physical machines are arranged in farms that perform roles in a network;

wherein the object to be placed on one of the plurality of hosts is at least one of;

a virtual machine (VM), a tenant or a database;

wherein the host is at least one of;

a physical machine, a database, a farm, a rack that holds physical machines, or a VM;

obtaining placement constraints for the object that comprise hard constraints and soft constraints, wherein application of a hard constraint eliminates one or more of the hosts from being selected on which to place the object, wherein soft constraints determine which hosts are more suitable than other hosts;

applying the hard constraints against hosts within the network that results in a pool of remaining hosts, wherein the hard constraints include high availability constraints, wherein applying the high availability constraints eliminates one or more hosts to ensure that objects having redundant roles within the network are placed on different hosts;

applying the soft constraints against the pool of remaining hosts that results in a pool of selection hosts;

automatically selecting a host on which to place the object from the pool of selection hosts; and

automatically deploying the object on the selected host.

11. The computer-readable storage medium of claim 10, wherein when the object is a VM and the plurality of hosts are physical machines, applying the hard constraints comprises applying two or more of the following constraints against each host:

host is located within the same network as the farm the VM is being deployed for;

host is of the correct type based on a role of the VM;

host is in the running state;

the host includes capacity to accommodate the VM;

or host does not already have a VM of the same role and farm.

12. The computer-readable storage medium of claim 10, wherein when the object is a tenant and the plurality of hosts are physical machines, the hard constraints comprise applying two or more of the following constraints against each host:

host is of a correct type;

host is in the running state;

host includes capacity to accommodate the database;

or host does not already have a database of the same type.

13. The computer-readable storage medium of claim 10, wherein the soft constraints are optimizations that when applied to the pool of remaining hosts spread the deployment of objects of a same type across different hosts.

14. The computer-readable storage medium of claim 10, wherein objects of a larger size are placed before objects having a smaller size.

15. The computer-readable storage medium of claim 10, wherein the hard constraints further include performance constraints.

16. A system for deploying objects on hosts for an online service, comprising:

a network comprising physical machines, racks for holding physical machines, virtual machines (VMs), servers, and databases;

wherein the servers are arranged in farms that each perform a role;

a processor and a computer-readable storage medium;

an operating environment stored on the computer-readable storage medium and executing on the processor; and

application code that is operative to;

determine when to place an object on one of a plurality of hosts;

wherein the object to be placed on one of the plurality of hosts is at least one of;

a virtual machine (VM), a tenant or a database;

wherein the host is at least one of;

a physical machine, a database, a farm, a rack or a VM;

obtain placement constraints for the object that comprise hard constraints and soft constraints, wherein application of a hard constraint eliminates one or more of the hosts from being selected on which to place the object, wherein soft constraints determine which hosts are more suitable than other hosts;

apply the hard constraints against hosts within the network that results in a pool of remaining hosts, wherein the hard constraints include high availability constraints, wherein applying the high availability constraints eliminates one or more hosts to ensure that objects having redundant roles within the network are placed on different hosts;

apply the soft constraints against the pool of remaining hosts that results in a pool of selection hosts;

automatically select a host on which to place the object from the pool of selection hosts; and

automatically deploy the object on the selected host.

17. The system of claim 16, wherein when the object is a VM and the plurality of hosts are physical machines, applying the hard constraints comprises applying two or more of the following constraints against each host:

host is located within the same network as the farm the VM is being deployed for;

host is of the correct type based on a role of the VM;

host is in the running state;

the host includes capacity to accommodate the VM;

or host does not already have a VM of the same role and farm.

18. The system of claim 16, wherein when the object is a database and the plurality of hosts are SQL server machines, the hard constraints comprise applying two or more of the following constraints against each host:

host is of a correct type;

host is in the running state;

or host includes capacity to accommodate the database.

19. The system of claim 16, wherein the soft constraints are optimizations that when applied to the pool of remaining hosts spread the deployment of objects of a same type across different hosts.

20. The system of claim 16, wherein the hard constraints further include performance constraints.