DATACENTER SYNCHRONIZATION

• US 20100235431A1
• Filed: 03/16/2009
• Published: 09/16/2010
• Est. Priority Date: 03/16/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Shah, Samir V., Poluri, Ravi Kiran R., Huang, Lin, Chen, Rui
• Granted Patent: US 8,291,036 B2
• Time in Patent Office: Days
• US Class Current: 709/203
• CPC Class Codes: G06F 11/2058 using more than 2 mirrored copies
• CPC Class Codes: G06F 11/2066 Optimisation of the communication load
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: G06F 11/2074 Asynchronous techniques
• CPC Class Codes: G06F 11/2076 Synchronous techniques
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation

Abstract

A datacenter infrastructure comprising a plurality of datacenters is configured to provide services to users. Data (e.g., application data) provided to respective datacenters from a user is synchronized between different datacenters comprised within “a cloud” according to a multilevel synchronization scheme. Respective levels of the multilevel synchronization operate to transfer different priority (user) data between the plurality of datacenters, wherein respective priorities are determined based upon the importance of the data in providing a service. For example, data used to perform a service is synchronized by a synchronization component facilitating a higher level of synchronization between datacenters, while data not routinely used to perform a service is synchronized by a lower priority synchronization component. Synchronization between datacenters allows a network address routing system to provide substantially uninterrupted routing of a user to an available or otherwise appropriate datacenter within the datacenter infrastructure, thereby providing increased service reliability and performance.

Claims

1. A computing environment configured to provide an improved reliability of services offered over a network, comprising:

a datacenter infrastructure comprising a plurality of datacenters, respective datacenters comprising a plurality of infrastructure components including a local master server configured to store user data; and

a multilevel synchronization component configured to synchronize data received from a user between the plurality of datacenters based upon a priority of the user data such that the user data is selectively routed to more than one of the plurality of datacenters, thereby enabling more than one of the plurality of datacenters to provide a service to the user.

2. The computing environment of claim 1, the multilevel synchronization component configured to perform a real time synchronization of a highest priority user data, based upon a received write command from a front end server comprised within respective datacenters to different datacenters.

3. The computing environment of claim 2, the multilevel synchronization component configured to perform an asynchronous replication of a medium priority user data comprised within the local master server, the medium priority user data able to tolerate inconsistencies between the plurality of datacenters while mitigating impact to the services offered to the user.

4. The computing environment of claim 3, the multilevel synchronization component configured to perform a synchronization between the plurality of datacenters performed at read time.

5. The computing environment of claim 1, respective datacenters comprising one or more secondary servers configured to store copies of the user data stored in respective local master servers.

6. The computing environment of claim 1, respective datacenters comprising a first IP address for a high priority traffic and a second IP address for a low priority traffic.

7. The computing environment of claim 6, wherein traffic received from the user is substantially balanced between the plurality of datacenters to reduce a traffic load received by a single datacenter.

8. The computing environment of claim 6, wherein upon failure of a connection between the user and one datacenter of the datacenter infrastructure, traffic received from the user are rerouted to one or more alternative datacenters comprised within the datacenter infrastructure.

9. The computing environment of claim 1, the datacenter infrastructure utilizing an anycast routing scheme.

10. The computing environment of claim 1, comprising a domain name system (DNS) configured provide substantially uninterrupted routing of the user to an available or otherwise appropriate datacenter within the datacenter infrastructure.

11. A method of multilevel synchronization across a plurality of datacenters configured to provide a service to a user, comprising:

receiving a user data at an active datacenter within one of the plurality of datacenters;

prioritizing the received user data according to an importance of the user data in providing the service to the user; and

performing a multilevel synchronization between the plurality of datacenters of a highest priority data, comprising;

making a real time synchronization request to write the user data to a local master server comprised within in respective datacenters;

checking for a response to the real time synchronization request to write; and

determining whether to proceed providing the service to the user based upon the response.

12. The method of claim 11, the multilevel synchronization comprising performing an asynchronous replication of a medium priority user data, which can tolerate inconsistencies between the plurality of datacenters while mitigating impact to the services offered to the user.

13. The method of claim 12, the multilevel synchronization comprising performing a synchronization between the plurality of datacenters performed at read time.

14. The method of claim 11, comprising assigning a high priority IP address and a low priority IP address to respective datacenters within the plurality of datacenters.

15. The method of claim 14, wherein upon a complete failure of a connection between the user and one of the plurality of datacenters, traffic received from the user is rerouted to the high priority IP address of another datacenter.

16. The method of claim 14, wherein upon a partial failure of a connection between the user and the first datacenter, traffic received from the user is rerouted to the low priority IP address of another datacenter.

17. The method of claim 11, the plurality of datacenters utilizing an anycast routing scheme so that the user data is received by a nearest active datacenter.

18. The method of claim 11, comprising writing user data stored within the local master server to one or more secondary servers comprised within respective datacenters.

19. The method of claim 18, wherein during a failure traffic of respective local master servers is rerouted to one of the one or more secondary servers comprised within respective datacenters.

20. A computing environment, comprising:

a datacenter infrastructure comprising a plurality of datacenters which utilizes an anycast routing scheme, respective datacenters comprising a plurality of infrastructure components including a local master server configured to store user data and one or more secondary servers configured to store copies of the user data stored in respective local master servers;

a multilevel synchronization component configured to synchronize data received from a user between the plurality of datacenters based upon a priority of the user data such that the user data is selectively routed to more than one of the plurality of datacenters, thereby enabling more than one of the plurality of datacenters to provide a service to the user, the multilevel synchronization component configured to perform;

a real time synchronization of a highest priority user data, the multilevel synchronization component implemented as write requests issued from a front end server comprised within respective datacenters to different datacenters;

an asynchronous replication of a medium priority user data comprised within the local master server, the medium priority user data able to tolerate inconsistencies between the plurality of datacenters while mitigating impact to the services offered to the user; and

a synchronization between the plurality of datacenters performed at read time; and

a domain name system (DNS) which assigns a first IP address for a high priority traffic and a second IP address for a low priority traffic to respective datacenters;

wherein upon failure of a connection between the user and one datacenter of the datacenter infrastructure, traffic received from the user are rerouted to one or more alternative datacenters comprised within the datacenter infrastructure.