Presentation is loading. Please wait.

Presentation is loading. Please wait.

Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 1 Reiner Ludwig Ericsson Research Eifel Response Algorithm: Ready for WGLC? Andrei Gurtov.

Similar presentations


Presentation on theme: "Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 1 Reiner Ludwig Ericsson Research Eifel Response Algorithm: Ready for WGLC? Andrei Gurtov."— Presentation transcript:

1 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 1 Reiner Ludwig Ericsson Research Eifel Response Algorithm: Ready for WGLC? Andrei Gurtov Sonera Corporation

2 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 2 Recall: The Problem (Spurious Timouts & Reordering) Spurious Retransmit of Entire Flight !!!

3 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 3 Recall: The Proposed Solution 1. Resolve Retrans. Ambiguity 2. Restore CWND / SSTHRESH & Update RTT Estimators 3. Resume “off the top”

4 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 4 To solve the problem you need … = Published RFC= I-D in WGLC= I-D in progress 3. Response E. Blanton & M. Allman Eifel Response Reverse CC State & Adapt DupThresh Reverse CC State & Adapt RTO 1. Signalling Timestamps RFC 1323 DSACK RFC 2883 “Heuristics”The RXT Flag or 2. Detection Eifel Detection RFC 3522 E. Blanton & M. Allman P. Sarolahti & M. Kojo or Spurious Timeouts & Reordering Only Spurious Timeouts Only Reordering

5 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 5 Eifel Response Limited to Spurious Timeouts Adapting DupThresh is Controversial –Don’t want to Change DupThresh for Occasional Reordering –When to start Adapting? How? –Do we need a DupThresh Estimator?  More Research Needed!  This Issues was Holding up Progress of the Eifel Response I-D  We will Exclude Adapting DupThresh in the Next Revision  Eifel Response will Only Work for Spurious Timeouts

6 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 6 Addressing the Remaining Open Issues Decaying of CC state after Detecting a Spurious Timeout –On 1 st Timeout (the one that later might be detected to be spurious): pipe_prev  max (FlightSize, ssthresh) –Every RTO (non-backed-off !!) counting from 1 st Timeout: pipe_prev  max (pipe_prev/2, SMSS) –On Detecting Spurious Timeout: cwnd  FlightSize + min (bytes_acked, IW) ssthresh  pipe_prev Adapting RTO in Response to Detecting a Spurious Timeout –With Timestamps: Reseed according to RFC2988: SRTT  R and RTTVAR  R/2 –Without Timestamps : RTTVAR  max (2 * RTTVAR, SRTT) and SRTT  2 * SRTT

7 Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 7 Question to the WG: Given that we Address the Remaining Issues as Outlined, is the Eifel Response I-D Ready for WGLC?


Download ppt "Reiner Ludwig – 57th IETF Meeting, July 2003Ericsson Research 1 Reiner Ludwig Ericsson Research Eifel Response Algorithm: Ready for WGLC? Andrei Gurtov."

Similar presentations


Ads by Google