Providing soft bandwidth guarantees using elastic TCP-based tunnels

• US 20050259689A1
• Filed: 04/01/2005
• Published: 11/24/2005
• Est. Priority Date: 04/01/2004
• Status: Abandoned Application

• Current Assignee: Trustees of Boston University
• Original Assignee: Trustees of Boston University
• Inventors: Matta, Abraham I., Bestavros, Azer
• Application Number:
  US11/096,735
• Publication Number:
  US 20050259689A1
• Time in Patent Office: Days
• US Class Current: 370/477
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/11 Identifying congestion
• CPC Class Codes: H04L 47/193 at the transport layer, e.g. TCP related
• CPC Class Codes: H04L 47/2425 for supporting services specification, e.g. SLA
• CPC Class Codes: H04L 69/08 Protocols for interworking; Protocol conversion
• CPC Class Codes: H04L 69/085 specially adapted for interworking of IP-based networks with other networks
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

Method and apparatus for providing enhanced utilization of an existing network of paths between nodes allocated to customer traffic where the paths also carry cross traffic. The system monitors the quality of the network bandwidth utilized by customer data flows over a set of managed paths in a time interval and allocates network resources to customers as a function of measured bandwidth and a desired target thereof by acquiring additional paths or abandoning existing paths. A scheduling function controls the use of the set of managed paths to more nearly achieve the desired quality of network bandwidth delivered to customer traffic.

Claims

1. A system for providing enhanced utilization of an existing network of paths between nodes allocated to customer traffic, said paths also carrying cross traffic, said system comprising:

means for monitoring average network bandwidth utilized by customer data flows over the paths in a time interval;

means for adjusting an allocation of bandwidth to customers as a function of measured average bandwidth and a desired bandwidth for customer use by acquiring or abandoning paths for users;

means for scheduling the use of the adjusted bandwidth paths for use by customers to more nearly achieve the desired bandwidth.

2. The system of claim 1 wherein said time interval is on the order of a few congestion epochs.

3. The system of claim 1 wherein said monitoring means determines an error function as the difference between desired bandwidth and measured average bandwidth.

4. The system of claim 1 wherein said monitoring means includes means for determining an error factor and said adjusting means allocates paths as a function of error factor.

5. The system of claim 4 wherein said adjusting means includes means for allocating paths as a function of said error factor either proportionally thereto or proportionally thereto and as an integral thereof.

6. The system of claim 4 wherein said error factor determining means determines error as a function of monitored flow history.

7. The system of claim 1 wherein said scheduling means includes one or more means for assigning different packets (or classes of packets) to different flows, selecting paths along which to open new flows, implementing admission control strategy for added tunnel demand, managing buffers associated with said scheduling means, and using redundant transmissions over multiple paths to meet specific constraints.

8. The system of claim 1 wherein said scheduling means includes means responsive to customer requirements including one or more functions supporting Virtual Private Network operations, Service Level Agreement, and QoS constraints.

9. The system of claim 1 further including means at a first node for encapsulating customer packet data with IP and TCP headers prior to sending over one or more said paths.

10. The system of claim 1 further including means associated with a node for stripping packet headings from data packets after passage through one or more paths.

11. A method for improving network data transfer using the system of any previous claim.

12. A method for providing enhanced utilization of an existing network of paths between nodes allocated to customer traffic, said paths also carrying cross traffic, said system comprising the steps of:

monitoring average network bandwidth utilized by customer data flows over the paths in a time interval;

adjusting an allocation of bandwidth to customers as a function of measured average bandwidth and a desired bandwidth for customer use by acquiring or abandoning paths for users;

scheduling the use of the adjusted bandwidth paths for use by customers to more nearly achieve the desired bandwidth.

13. The method of claim 12 wherein said time interval is on the order of a few congestion epochs.

14. The method of claim 12 wherein said monitoring step determines an error function as the difference between desired bandwidth and measured average bandwidth.

15. The method of claim 12 wherein said monitoring step determines an error factor representing difference between desired bandwidth and measured bandwidth and said adjusting step allocates paths as a function of error factor.

16. The method of claim 15 wherein said adjusting step allocates paths as a function of said error factor either proportionally thereto or proportionally thereto and as an integral thereof.

17. The method of claim 15 wherein said error factor determining step determines error as a function of monitored flow history.

18. The method of claim 12 wherein said scheduling step is operative for one or more of assigning different packets (or classes of packets) to different flows, selecting paths along which to open new flows, implementing admission control strategy for added tunnel demand, managing buffers associated with said scheduling means, and using redundant transmissions over multiple paths to meet specific constraints.

19. The method of claim 12 wherein said scheduling step is responsive to customer requirements including one or more functions supporting Virtual Private Network operations, Service Level Agreement, and QoS constraints.