Title: Diffusion Mechanisms for Active Queue Management
1Diffusion Mechanisms for Active Queue Management
Rafael Nunez nunez_at_ece.udel.edu
Gonzalo Arce arce_at_ece.udel.edu
- Department of Electrical and Computer Engineering
- University of Delaware
- May 19th / 2004
2Diffusion Mechanisms for Active Queue Management
- Introduction
- Diffusion Early Marking
- Performance
- Conclusions and Future Work
3The Internet Today
4Congestion
- Desirable control distributed, simple, stable
and fair.
5Problems with Tail Dropping
- Penalizes bursty traffic
- Discriminates against large propagation delay
connections. - Global synchronization.
6Active Queue Management (AQM)
- Random Early Detection (Floyd and Jacobson, 1993)
- Router becomes active in congestion control.
- RED has been deployed in some Cisco routers.
7Random Early Detection (RED)
- Random packet drops in queue.
- Drop probability based on average queue
- Four parameters
- qmin
- qmax
- Pmax
- wq
- (overparameterized)
8Queue Behavior in RED
9Queue Behavior in RED (2)
- 20 new flows every 20 seconds
10Other AQMs Schemes
- Adaptive RED, REM, GREEN, BLUE,
- Problems
- Over-parameterization
- Not easy to implement in routers
- Not much better performance than drop tail
11Diffusion Mechanisms for Active Queue Management
- Introduction
- Diffusion Early Marking
- Performance
- Conclusions and Future Work
v
12Diffusion Mechanisms for AQM
- Instantaneous queue size.
- Better packet marking strategy.
- Simplified parameters.
13Probability of Marking a Packet
- Gentle RED function closely follows
(A)
14Evolution of the Congestion Window
(B)
15Traffic in the Network
- Congestion Window Packets In The Pipe Packets
In The Queue - Or
(C)
- From (A), (B), (C), and knowing that
where
16Probability Function
17Error Diffusion
- Packet marking is analogous to halftoning
- Convert a continuous gray-scale image into black
or white dots - Packet marking reduces to quantization
- Error diffusion The error between input
(continuous) and output (discrete) is
incorporated in subsequent outputs.
18Diffusion Mechanism
19Diffusion Mechanism
20Diffusion Mechanism
21Diffusion Mechanism
22Diffusion Mechanism
23Diffusion Mechanism
24Diffusion Mechanism
25Diffusion Mechanism
26Diffusion Mechanism
27Diffusion Mechanism
28Diffusion Mechanism
29Diffusion Mechanism
30Algorithm Summary
- Diffusion Early Marking decides whether to mark a
packet or not as
Where
Remember
M2, b12/3, b21/3
31Diffusion Mechanisms for Active Queue Management
- Introduction
- Diffusion Early Marking
- Performance
- Conclusions and Future Work
v
v
32Dropping Packets
33Window Size
2 Flows
10 Flows
34Stability of the Queue
- 100 long lived connections (TCP/Reno, FTP)
- Desired queue size 30 packets
35Changing the number of flows
- 20 new flows every 20 seconds
36Diffusion Mechanisms for Active Queue Management
- Introduction
- Diffusion Early Marking
- Performance
- Conclusions and Future Work
v
v
v
37Conclusions and Future Work
- Queue length stabilized and controlled without
adjusting parameters. - Diffusion mechanism improves the behavior of the
proposed AQM scheme. - Future Work
- Estimation of parameters
- Analyze more traffic scenarios
- Compare with other AQMs
- Use diffusion mechanism in other AQMs
38Diffusion Mechanisms for Active Queue Management
- Introduction
- Diffusion Early Marking
- Parameter Estimation
- Conclusions and Future Work
v
v
v
v