Presentation is loading. Please wait.

Presentation is loading. Please wait.

Benchmarking Network-layer Traffic Control Mechanisms

Similar presentations


Presentation on theme: "Benchmarking Network-layer Traffic Control Mechanisms"— Presentation transcript:

1 Benchmarking Network-layer Traffic Control Mechanisms
Jerry Perser David Newman Sumit Khurana Shobha Erramilli Scott Poretsky 57th IETF Vienna, Austria

2 Changes to dsmterm-07 Added initial copyright statement to the first page. Took the first paragraph of Introduction and created the Abstract. Reworded the discussion of Forwarding Congestion. Changed Conforming and Nonconforming to Conforming Packet and Nonconforming Packet. Added a discussion to Duplicate Packet. We only had a definition before. Added RFC2119 to the reference section.

3 Methodology - Goals Purpose: provide methodologies to benchmark devices "capable of delivering the specific packet forwarding treatment indicated by the DS field value" [RFC2474]. Benchmark: Device’s ability to meet configured PHBs (the Expected Vector), not QoS mechanisms. Possible to compare performance of routers implementing different mechanisms. Goals consistent with RFC2474: “A vendor MAY use any [QoS] algorithm that satisfies the definition of the PHB according to the standard. The node's capabilities and its particular configuration determine the different ways that packets can be treated."

4 Methodology – General Procedure
RFC 2475: “The most simple example of a PHB is one which guarantees a minimal bandwidth allocation of X% of a link (over some reasonable time interval) to a behavior aggregate. This PHB can be fairly easily measured under a variety of competing traffic conditions.” Test cases to benchmark PHBs require: Selection Codepoint Set Selection and configuration of Expected Vectors Sourcing of Offered Vector Measurement of all Output Vectors

5 Methodology – Test Cases
EF PHB EF PHB Receives Absolute Priority EF PHB Forwarding Vector Matches the Expected Vector EF PHB Rate-Limiting (with Packet Drops or Remarking) EF PHB Does not Waste Bandwidth AF PHB AF PHB Forwarding Vectors Match the Expected Vectors AF PHBs are Fairly Treated with EF PHB AF PHB Rate-Limiting (with Packet Drops or Remarking) AF PHB Congestion Avoidance CS PHB CS PHB is Backwards Compatible with IP Precedence CS PHB Forwarding Vectors Match the Expected Vectors CS PHBs are Fairly Treated with EF and AF PHB CS PHB Rate-Limiting (with Packet Drops or Remarking) CS PHB Congestion Avoidance Default PHB Unmarked Packets Receive Best-Effort Forwarding Unmarked Packets are Tail-Dropped Configured Expected Vector DUT Offered Vector 1 Test Equip Forwarding Vector Offered Vector 2

6 Any Questions ?


Download ppt "Benchmarking Network-layer Traffic Control Mechanisms"

Similar presentations


Ads by Google