UNCOVERING THE DIFFERENCES IN BACKBONE NETWORKS

• US 20090086741A1
• Filed: 10/02/2007
• Published: 04/02/2009
• Est. Priority Date: 10/02/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Mahajan, Ratul, Zhang, Ming
• Granted Patent: US 7,817,547 B2
• Time in Patent Office: Days
• US Class Current: 370/401
• CPC Class Codes: H04L 41/5009 Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
• CPC Class Codes: H04L 41/5077 wherein the managed service relates to simple transport services, i.e. providing only network infrastructure
• CPC Class Codes: H04L 43/12 Network monitoring probes
• CPC Class Codes: H04L 43/55 Testing of service level quality, e.g. simulating service usage

Abstract

The claimed subject matter provides systems and/or methods that uncovers detailed differences in the performance of different Internet Service Provider ISP) networks. The system can include components, devices, and/or processes that utilize a list of target Internet Service Provider (ISP) networks to create a list of Internet Protocol (IP) addresses to interrogate, distribute the list of addresses to measurement nodes and obtains in return probing results from the measurement nodes, the probing results employed thereafter to provide comparative performance metrics associated with the target Internet Service Provider (ISP) networks.

Claims

1. A system implemented on a set of machines that uncovers detailed differences in performance of Internet Service Provider (ISP) networks, comprising:

a controller that utilizes a list of target networks to create a list of destination Internet Protocol (IP) addresses to interrogate, the controller propagates the list of destination addresses to measurement nodes and obtains in return probing results from the measurement nodes, the probing results employed by the controller to provide a comparative performance of the target networks.

2. The system of claim 1, the controller employs Border Gateway Protocol atoms to reduce sustained bandwidth usage.

3. The system of claim 2, atoms include groups of IP (Internet Protocol) prefixes with identical routing policies such that paths to included IP (Internet Protocol) prefixes are identical.

4. The system of claim 1, the controller utilizes a routing view generated by a measurement node to restrict probing to of Internet Service Provider (ISP) networks of interest.

5. The system of claim 1, the controller adds probes to a measurement set until all elements are covered at least once.

6. The system of claim 5, elements include paths to the destination Internet Protocol (IP) addresses, paths internal to Internet Service Provider (ISP) networks, or measurement nodes.

7. The system of claim 5, the controller on completion of each processing cycle adds probes that cover a majority of as yet uncovered elements.

8. The system of claim 1, the controller checks for occurrences of errors in DNS (Domain Name Service) names or majority voting through utilization of accumulated traceroute data.

9. The system of claim 1, the controller employs single-ended measurements to infer path latency.

10. The system of claim 1, the controller discards traceroutes for which forward or reverse hop lengths to an IP (Internet Protocol) address of interest differs by more than three.

11. The system of claim 10, the forward hop length determined from a traceroute measurement.

12. The system of claim 10, the reverse hop length inferred using a remaining Time To Live (TTL) value associated with a probe response.

13. The system of claim 1, the controller detects overload events by determining a variance in a round trip time from a measurement node to an IP (Internet Protocol) address where a probe enters the target network.

14. The system of claim 1, further includes a data analyzer that extracts information associated with subpaths of interest from raw measurements returned by the measurement nodes.

15. The system of claim 14, the data analyzer utilizes the extracted information associated with the subpaths to provide latency samples of the subpaths.

16. The system of claim 15, the data analyzer groups the latency samples of the subpaths by source of where a measurement is taken.

17. The system of claim 15, the data analyzer discards latency samples based on a determination of asymmetrical paths, or aggregates latency samples into an estimated duration for subpaths or interest.

18. A method implemented on a set of machines that reveals detailed differences in performance of target Internet Service Provider (ISP) networks, comprising:

developing a list of Internet Protocol (IP) addresses to interrogate; and

distributing the list of addresses to measurement nodes and retrieving measurement results from the measurement nodes, the measurement results employed to provide a comparative performance of the target Internet Service Provider (ISP) networks.

19. The method of claim 18, the comparative performance employed by applications to distribute processes or sub-processes for execution on destinations associated with networks with high comparative performance.

20. A system that provides detailed differences in performance of Internet Service Provider (ISP) networks, comprising:

means for producing a list of destination Internet Protocol (IP) addresses to interrogate;

means for distributing the list of destination addresses for interrogation;

means for recovering raw probing results associated with the destination addresses from the means for interrogation; and

means for converting the raw probing results to provide a comparative performance of the target Internet Service Provider (ISP) networks.