1 / 15

Resource/Accuracy Tradeoffs in Software-Defined Measurement

HotSDN’13. Resource/Accuracy Tradeoffs in Software-Defined Measurement. Masoud Moshref, Minlan Yu, Ramesh Govindan. Management policies need measurement. Traffic Engineering: Find elephant flows to route [ Hedera ] Estimate rack-to-rack traffic matrix [ MicroTE ]. Accounting:

abiola
Download Presentation

Resource/Accuracy Tradeoffs in Software-Defined Measurement

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. HotSDN’13 Resource/Accuracy Tradeoffs in Software-Defined Measurement Masoud Moshref, Minlan Yu, Ramesh Govindan

  2. Management policies need measurement • Traffic Engineering: • Find elephant flows to route [Hedera] • Estimate rack-to-rack traffic matrix [MicroTE] Accounting: Tiered pricing based on network usage • Troubleshooting: • Find network bottlenecks of an application • Incast problem

  3. Software Defined Measurement Controller Traffic Engineering 1 2 1 Configure resources Fetch statistics (Re)Configure resources 3

  4. Challenges • Limited resources • Limited CPU/memory in switches • Limited control network bandwidth • Complexity of different primitives in switches • Counting by prefix matching (TCAM) • Counting based on hashes • Programming • Complexity of different primitives in switches • Counting by prefix matching (TCAM) • Counting based on hashes • Programming • Different measurements • Resource usage depends on measurement

  5. Using Different Primitives for Measurement Comparing primitives in performing a measurement task Example: Finding heavy hitter IP when traffic from 10.0.1.0/24 goes beyond a threshold

  6. Counting TCAM matches Counters Filters 18 0 1 1 0 * * * * • Large time-scale measurement • Measure-install loop latency • Controller link bandwidth • Slowly varying traffic Monitor 10.0.1.0/24 When goes beyond threshold, iteratively search (e.g. binary search) Controller - 1 2 3 10.0.1.0/24 10.0.1.0/25 10.0.1.128/25 10.0.1.128/26 Install Fetch . . .

  7. Sketches using hash based counters Src:10.0.1.5, Size:2 3 • Uses cheaper resources (SRAM) • Not dependent on traffic history • Use for varying traffic • Still history may help in large time scale to optimize parameters Fetch counters Approximate traffic for each IP Controller H 2 1 5 1 0 0 1 0 5 3 1 0 2

  8. Programming switches • Use more complex data structures/commands • Heap for finding large flows 10.0.1.5 90 10.0.1.1 25 10.0.1.3 50 10.0.1.12 17 10.0.1.10 3 • Uses CPU resources • Can it keep up with line rate?

  9. Summary of tradeoffs of primitives

  10. Case Study: Hierarchical Heavy Hitter Detection

  11. Hierarchical Heavy Hitters 36 *** 26 10 0** 1** 12 14 5 5 00* 01* 10* 11* 111 001 011 101 5 7 12 2 0 5 2 3 010 110 000 100

  12. Hierarchical Heavy Hitters • Longest IP prefixes • Contribute a large amount of traffic • After excluding any HHH descendants in the prefix tree 36 Threshold=10 *** 26 10 0** 1** 12 14 5 5 00* 01* 10* 11* 111 001 011 101 5 7 12 2 0 5 2 3 010 110 000 100

  13. Algorithms for single switch HHH detection • Prefix-based counting • Proposed an iterative algorithm at controller • Given TCAM capacity, monitor prefixes that maximize accuracy 5 26 5 26 3 - 2 - 3 13 2 13 • Hash-based counting • Use Hierarchical Count-Min sketch • Approximate traffic in all levels using sketches • Efficiently traverse IP tree at controller to find HHHs 30 30 26 Approximate Traverse tree 20 9 20 18 9 8

  14. Evaluation Normalize switch cost 80 SRAM 1 TCAM SRAM counters TCAM counters Sketches have better accuracy for small threshold (longer prefixes, higher variability) Max-Cover saves bandwidth for large number of counters and large thresholds

  15. Future work Controller • Use joint information • Guarantee accuracy Traffic Engineering Accounting North-bound Anomaly Detection Software Defined Measurement • Select primitive • Assign switch resources • Run global measurement South-bound

More Related