Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model...

18
Incorporating Intra-flow Dependencies and Inter-flow Correlations for Traffic Matrix Prediction Presenter: Kaihui Gao Tsinghua University

Transcript of Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model...

Page 1: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Incorporating Intra-flow Dependencies and Inter-flow Correlations for Traffic

Matrix Prediction

Presenter: Kaihui Gao

Tsinghua University

Page 2: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Catalog

1. Background

2. Related Works

3. Motivation

4. Data Analysis of Traffic Matrix

5. Challenge

6. Solution

7. Evaluation

Page 3: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Background

• Accurate TM prediction plays an important role in the areas of traffic engineering (TE), capacity planning, resource allocation, congestion mitigation, and failure/anomaly detection in computer networks.

• For network traffic, as a kind of spatial-temporal series, its forecasting has to face similar key challenges of modeling sequential information as the traditional time series forecasting problem. Furthermore, accurate forecasting requires addressing the challenges of modeling the complex and underlying spatial correlation.

Page 4: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Problem Formulation

• TM: Traffic Demand Matrix• (TMt+1 ,TMt+2,…, TMt+q ) = f(TMt,TMt-1,…,TMt-w+1)

12

34

5

67

8

9

10

11

12

Page 5: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Related Works

• ARMA, ARIMA (Auto-regressive Integrated Moving Average)• linear structure cannot model the nonlinear patterns

• SVR (Support Vector Regression)• significant effort is needed to tune its parameters

• Neural Networks (NNs) based approaches• RNN: LSTM, GRU

Page 6: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

State-of-the-art: RNN

TMt-w+1

TMt-w+2

TMt-1

TMt

TMt+1

Page 7: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Motivation

server

client1

client2

request

request

response

response

• Popular service ——> positive correlations• Bandwidth contention ——> negative correlations

Page 8: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Spatial correlation analysis

Pearson Correlation Coefficient:

Observation 1: There exist abundant high correlations between flows in real WAN.

Page 9: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Spatial correlation analysis

Observation 2: Most strongly-correlated flow pairs sharethe same source or destination.

Page 10: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

How to capture spatial correlations

��

• Explicitly• high computational overhead• curse of dimension• time-varying

• Implicitly• learning based• convolutional neural network

Page 11: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Solution

��

•We propose a novel deep learning framework combined CNN and RNN to predicting future traffic matrix.

• Combined of• CNN capture the spatial correlations of every flow pair.• RNN model the temporal variation of spatial correlations

and traffic.

Page 12: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Overview

��

Page 13: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Evaluation

��

• Datasets: Abilene and GEANT• training set : test set = 8 : 2• Performance Metric: •Mean Square Error•Mean Absolute Error

•Baseline: ARIMA, SVR, LSTM, CNN

Page 14: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Predicting Next Single TM

��

Page 15: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Predicting Next Multiple TMs

��

Page 16: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Traffic Engineering based on TM Prediction

��

• TE problem: given a network with edge capacities and a set of traffic demands of source-destination pairs, we try to find an assignment of demands to paths (an ordered sequence of edges) that optimizes some objectives. A widely used TE objective is minimizing the maximum link utilization (MLU).

• Multicommodity Flow (MCF) problem

• Baseline�• MCF based on real TMs• MCF based on predicated TMs by CRNN• MCF based on predicated TMs by LSTM• shortest path routing (SP)• Oblivious routing (OR)

Page 17: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

Performance comparison

��

Page 18: Incorporating Intra-flow Dependencies and Inter …2020-slides.pdf• linear structure cannot model the nonlinear patterns •SVR (Support Vector Regression) •significant effort

!