18
A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Embed Size (px)

Citation preview

Page 1: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

A Practical Traffic Management System for Integrated

LTE-WiFi Networks

Zhuoran Li11/4/2015

Page 2: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Overview

• Problem:– With the data traffic increasing fast, operators

start to deploy Wi-Fi for additional capacity, therefore how to manage the network interface for user flows is important.

Page 3: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Drawbacks of current solution

• Naïve policies: in most cases, the connection managers are on the user device. (User decides to connect to WiFi AP or basestation.)

• Static decision: interface selection isn’t adaptive to dynamic conditions of wireless networks.

• Coarse-grained policies: interface selections based on the user-level. However, application-level based selection will be better.

Page 4: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Their Idea: ATOM

LTE Core-Network

WiFi Gateway

Traffic Manager

Maps user flows to appropriate network(LTE/WiFi) Centralized management -> Efficient use of network

resources Reduces backhaul costs -> Facilitates dynamic traffic

mgmt Operates for each LTE cell -> Scalable Standards agnostic -> Easily Deployable

PDN-gateway

MME

Serving-gateway

Network Interface

Assignment

Switching Service

Page 5: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Components Network Interface Assignment Algorithm

(NIA)– Goal: Dynamically maps user traffic flows to

appropriate LTE basestation or WiFi AP

• Interface switching service (ISS)– Goal: Switch current user flows from WiFi AP to

LTE or vice versa based on decisions from NIA

Page 6: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Component 1: NIA

6

Page 7: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Consider an LTE cell and multiple WiFi APs in its coverage area

• Assign basestation/ AP to each flow– Maximize sum of users flows’ QoE

• QoE captured using “utility”– Weighted PF provides differential QoE

• Pricing function supports 2 models– Based on data usage– Based on price/byte

Problem Formulation

7

k

max kU

())1log( PtwU kk

Weight Throughput Network Pricing

Page 8: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Network Interface Assignment (NIA)

• Problem is NP-Hard– Including the simplest topology of an LTE cell and a WiFi AP

• NIA is a two-step greedy heuristic– Considers each AP-basestation in isolation– Fixes assignment for AP that maximized incremental utility– Iterate till all hotspots are covered– Complexity is O(B2N), where N = # flows, B = # APs

8

Page 9: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

NIA algorithm

Page 10: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Component 2: Interface Switching Service

Page 11: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Interface Switching Service (ISS)

Page 12: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

ISS

• They focus on the HTTP-based traffic flows.– HTTP-based downloads: go through core-network,

connecting with the HTTP Proxy in the Core-Network.

– HTTP-based video streaming and browsing: for this traffic, they design it go directly to the Internet.

– Non HTTP-traffic: go through the core network.

Page 13: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Evaluation

In WiFi AP1, 4 users steamYouTube videos of average bit-rate 1.5 Mbps while 1 user downloads a large-file. In WiFi AP2, 2 users stream YouTube videos of average bit-rate 1.5 Mbps while 2 users download large-files. Both the LTE users stream YouTube videos of average bit-rate 1.5 Mbps. We compare the throughput of WiFi users for the case with and without ATOM (NO-ATOM).

With ATOM, User #3 and user #6 are moved to LTE-basestation.

Page 14: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Evaluation

There are 4 users streaming YouTube videos of average bit-rate of 1.5 Mbps. Users#1, 2 and 3 are within WiFi coverage while User#4 can only access the eNodeB. We introduce 2WiFi users with background traffic at around 30 seconds into the experiment such that the flows are active for about 40 seconds

Page 15: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Benchmarking the ISSMeasured the time taken for flows to switch using ISS:

• HTTP based video streaming flows• Hulu (uses HTTP-DASH) v/s Youtube(uses HTTP-PD)

Insight: Switching time improves with DASH streaming• DASH flows use smaller chunk sizes to ensure adaptive-ness

to changing network conditions

Page 16: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

My understanding

• Pros:– The most important contribution of this work is

that they build a real system. For interface selection, a lot of searcher provides a lot of algorithm, however, this is the first one to build the real system.

– To verify their system, they build a small testbed and also did a large-scale simulation. The combination of real testbed and simulation can convince people about their system.

Page 17: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

My understanding

• My questions:– Is that possible for a mobile device to send data

though both cellular and wifi? If that is possible, why a user have to choose one interface not transferred at both?

– They didn’t say clearly if they make the HTTP-based video streaming go directly to the Internet, how the charge this data?

Page 18: A Practical Traffic Management System for Integrated LTE-WiFi Networks Zhuoran Li 11/4/2015

Thanks!