INFOCOM 2013 – Torino, Italy Content-centric wireless networks with limited buffers: when mobility hurts Giusi Alfano, Politecnico di Torino, Italy Michele.

Slides:



Advertisements
Similar presentations
Xia Zhou*, Stratis Ioannidis ♯, and Laurent Massoulié + * University of California, Santa Barbara ♯ Technicolor Research Lab, Palo Alto + Technicolor Research.
Advertisements

1 On the Capacity of Ad Hoc Wireless Networks Under General Node Mobility Michele Garetto – Università di Torino Paolo Giaccone - Politecnico di Torino.
Impact of Interference on Multi-hop Wireless Network Performance Kamal Jain, Jitu Padhye, Venkat Padmanabhan and Lili Qiu Microsoft Research Redmond.
Mobility Increase the Capacity of Ad-hoc Wireless Network Matthias Gossglauser / David Tse Infocom 2001.
August 16, 2014 Modeling the Performance of Wireless Sensor Networks Carla Fabiana Chiasserini Michele Garetto Telecommunication Networks Group Politecnico.
Improving TCP over Wireless by Selectively Protecting Packet Transmissions Carla F. Chiasserini Michele Garetto Michela Meo Dipartimento di Elettronica.
The Capacity of Wireless Ad Hoc Networks
A Layered Hybrid ARQ Scheme for Scalable Video Multicast over Wireless Networks Zhengye Liu, Joint work with Zhenyu Wu.
Peering in Infrastructure Ad hoc Networks Mentor : Linhai He Group : Matulya Bansal Sanjeev Kohli EE 228a Course Project.
NCKU CSIE CIAL1 Principles and Protocols for Power Control in Wireless Ad Hoc Networks Authors: Vikas Kawadia and P. R. Kumar Publisher: IEEE JOURNAL ON.
Mobility Increases Capacity In Ad-Hoc Wireless Networks Lecture 17 October 28, 2004 EENG 460a / CPSC 436 / ENAS 960 Networked Embedded Systems & Sensor.
1 Using Multiple Channels and Spatial Backoff to Improve Wireless Network Performance Nitin Vaidya University of Illinois at Urbana-Champaign
CS401 presentation1 Effective Replica Allocation in Ad Hoc Networks for Improving Data Accessibility Takahiro Hara Presented by Mingsheng Peng (Proc. IEEE.
Mobile Ad hoc Networks COE 549 Delay and Capacity Tradeoffs II Tarek Sheltami KFUPM CCSE COE 8/6/20151.
Network Coding vs. Erasure Coding: Reliable Multicast in MANETs Atsushi Fujimura*, Soon Y. Oh, and Mario Gerla *NEC Corporation University of California,
Fundamental Lower Bound for Node Buffer Size in Intermittently Connected Wireless Networks Yuanzhong Xu, Xinbing Wang Shanghai Jiao Tong University, China.
A Simple and Effective Cross Layer Networking System for Mobile Ad Hoc Networks Wing Ho Yuen, Heung-no Lee and Timothy Andersen.
EE360 PRESENTATION On “Mobility Increases the Capacity of Ad-hoc Wireless Networks” By Matthias Grossglauser, David Tse IEEE INFOCOM 2001 Chris Lee 02/07/2014.
1 Core-PC: A Class of Correlative Power Control Algorithms for Single Channel Mobile Ad Hoc Networks Jun Zhang and Brahim Bensaou The Hong Kong University.
Improving Capacity and Flexibility of Wireless Mesh Networks by Interface Switching Yunxia Feng, Minglu Li and Min-You Wu Presented by: Yunxia Feng Dept.
Computer Networks Performance Metrics. Performance Metrics Outline Generic Performance Metrics Network performance Measures Components of Hop and End-to-End.
E VALUATION OF F AIRNESS IN ICN X. De Foy, JC. Zuniga, B. Balazinski InterDigital
1 Mobility Increases the Capacity of Ad-hoc Wireless Networks Matthias Grossglauser, David Tse IEEE Infocom 2001 (Best paper award) Oct 21, 2004 Som C.
Converge-Cast: On the Capacity and Delay Tradeoffs Xinbing Wang Luoyi Fu Xiaohua Tian Qiuyu Peng Xiaoying Gan Hui Yu Jing Liu Department of Electronic.
1 Performance Analysis of the Distributed Coordination Function under Sporadic Traffic joint work with C.-F. Chiasserini (Politecnico di Torino)
S Master’s thesis seminar 8th August 2006 QUALITY OF SERVICE AWARE ROUTING PROTOCOLS IN MOBILE AD HOC NETWORKS Thesis Author: Shan Gong Supervisor:Sven-Gustav.
TOPOLOGY MANAGEMENT IN COGMESH: A CLUSTER-BASED COGNITIVE RADIO MESH NETWORK Tao Chen; Honggang Zhang; Maggio, G.M.; Chlamtac, I.; Communications, 2007.
MAIN RESULT: Depending on path loss and the scaling of area relative to number of nodes, a novel hybrid scheme is required to achieve capacity, where multihop.
Tufts Wireless Laboratory School Of Engineering Tufts University Paper Review “An Energy Efficient Multipath Routing Protocol for Wireless Sensor Networks”,
Content caching and scheduling in wireless networks with elastic and inelastic traffic Group-VI 09CS CS CS30020 Performance Modelling in Computer.
SERENA: SchEduling RoutEr Nodes Activity in wireless ad hoc and sensor networks Pascale Minet and Saoucene Mahfoudh INRIA, Rocquencourt Le Chesnay.
Courtesy Piggybacking: Supporting Differentiated Services in Multihop Mobile Ad Hoc Networks Wei LiuXiang Chen Yuguang Fang WING Dept. of ECE University.
Puzzle You have 2 glass marbles Building with 100 floors
Routing Protocols to Maximize Battery Efficiency
MAC Protocols for Sensor Networks
Impact of Interference on Multi-hop Wireless Network Performance
2010 IEEE Global Telecommunications Conference (GLOBECOM 2010)
Delay-Tolerant Networks (DTNs)
Data Center Network Architectures
Presented by Tae-Seok Kim
David S. L. Wei Joint Work with Alex Chia-Chun Hsu and C.-C. Jay Kuo
Group Multicast Capacity in Large Scale Wireless Networks
Ad-hoc Networks.
Notes Onur Ascigil, Vasilis Sourlas, Ioannis Psaras, and George Pavlou
Computing and Compressive Sensing in Wireless Sensor Networks
Delay Efficient Wireless Networking
A Study of Group-Tree Matching in Large Scale Group Communications
Mohammad Malli Chadi Barakat, Walid Dabbous Alcatel meeting
GeoTORA: A Protocol for Geocasting in Mobile Ad Hoc Networks
Routing in Wireless Ad Hoc Networks by Analogy to Electrostatic Theory
Introduction Secondary Users (SUs) Primary Users (PUs)
Scheduling in Wireless Communication Systems
Computer Network Performance Measures
High Throughput Route Selection in Multi-Rate Ad Hoc Wireless Networks
Computer Network Performance Measures
Throughput-Optimal Broadcast in Dynamic Wireless Networks
A short lesson on wireless connectivity…
Effective Replica Allocation
The Impact of Multihop Wireless Channel on TCP Performance
Effectiveness of the 2-hop Routing Strategy in MANETS
Capacity of Ad Hoc Networks
IFIP – Performance 2007 A Modeling Framework to Understand the Tussle between ISPs and Peer-to-Peer File Sharing Users Michele Garetto - unito.
Javad Ghaderi, Tianxiong Ji and R. Srikant
Pradeep Kyasanur Nitin H. Vaidya Presented by Chen, Chun-cheng
Gaurav Sharma,Ravi Mazumdar,Ness Shroff
Subject Name: Adhoc Networks Subject Code: 10CS841
of the IEEE Distributed Coordination Function
Optimal Control for Generalized Network-Flow Problems
Satellite Packet Communications A UNIT -V Satellite Packet Communications.
Presentation transcript:

INFOCOM 2013 – Torino, Italy Content-centric wireless networks with limited buffers: when mobility hurts Giusi Alfano, Politecnico di Torino, Italy Michele Garetto, Università di Torino, Italy Emilio Leonardi, Politecnico di Torino, Italy

Outline Motivation Previous work System assumptions Main results Possible extensions

Why another paper on scaling laws of wireless networks? The way users search and retrieve data is changing: from host-to-host paradigm to host-to-content paradigm Contents are usually replicated to save bandwidth and improve QoS e.g.: content-delivery networks (CDN) The dominant traffic is going to be anycast, and not unicast

Existing scaling laws for wireless networks The traditional way to start: ″consider N (unicast) flows randomly established among the nodes…″ e.g.: Gupta-Kumar (1999) Grossglauser-Tse (2001) A lot of work also on multicast… Only few papers on anycast (and associated caching strategy) all of them assume static nodes

Joint replication and delivery problem Previous work S. Gitzenis, G.S. Paschos, L. Tassiulas, “Asymptotic laws for content replication and delivery in wireless networks” (INFOCOM ′12)   We cache #123 ! grid I want content #123 ! Here it is ! Joint replication and delivery problem

System assumptions 𝑁 nodes moving in a square (of area 1) 𝑀 contents (𝑀= 𝑁 𝛽 , 0 ≤𝛽≤1) Zipf-like popularity: 𝑝 𝑖 = 𝐻 𝑖 𝛼 , 𝛼≥0 𝐾 cache size at each node. 𝐾 finite!! measured in number of equal-size contents (same scaling laws if bounded ratio maxsize/minsize)

Communication model Nodes can either use: unique transmission range 𝑅 content-dependent transmission range: 𝑅 𝑚 : transmision range of content 𝑚 (with power control to compensate attenuation) Interference is taken into account by: Physical model (𝑆𝐼𝑁𝑅> 𝜎) …which is shown to be equivalent to a: Generalized protocol model (also with variable tx-ranges)

Traffic model Each transmission between two nodes allows to exchange an entire content (same scaling laws if you exchange only a segment of a content, each content split in bounded number of segments – no fluid limit) Each node cycles these steps: requests a random content (Zipf law) waits until content is received waits an additional idle time (to allow throughput-delay trade-offs) requests another content… Which implies: at most one pending content request per node (same scaling laws if bounded number of parallel requests)

Mobility model We first consider for simplicity: reshuffling mobility model (i.i.d.): (new random topology generated at each time slot) then we extend the analysis to: random walk (each node displaced by flight size 𝐹 from slot to slot) (but no communication while moving) Flight size 𝐹 is varied from 1/ 𝑁 (quasi-static network) to 1 (similar to i.i.d.)

Contents’ replication As in previous work, we assume caches of nodes are pre-populated by the system As consequence of the fact that we consider static set of contents with constant popularity No run-time optimizations (cache replacements induced by traffic) We leave to future work: dynamic set of contents with varying popularity

Main results Performance metrics: 𝜆 : per-node throughput (in contents/slot) 𝐷 : average content transfer delay We are interested in trade-offs between 𝜆 and 𝐷 Things we can play with: number of replicas for each content transmission range(s) idle time between successive content requests …and of course a communication scheme !…

For the reshuffling (i.i.d.) mobility: Main results For the reshuffling (i.i.d.) mobility: Using unique transmission range 𝑅: 𝛼>2 : best possible performance 𝜆=Θ 1 , 𝐷=Θ 1 1<𝛼<2 : 𝐷=Θ 𝜆 𝑀 2−𝛼 𝛼<1: 𝐷=Θ 𝜆 𝑀 Using content-adaptable transmission ranges 𝑅 𝑚 : 𝛼>3/2 : best possible performance 𝜆=Θ 1 , 𝐷=Θ 1 1<𝛼<3/2 : 𝐷=Θ 𝜆 𝑀 3−2𝛼 𝑇ℎ𝑒𝑠𝑒 𝑎𝑟𝑒 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑟𝑒𝑠𝑢𝑙𝑡𝑠 𝑎𝑐ℎ𝑖𝑒𝑣𝑎𝑏𝑙𝑒 𝑖𝑛 𝑎 𝑠𝑡𝑎𝑡𝑖𝑐 𝑔𝑟𝑖𝑑 𝑛𝑒𝑡𝑤𝑜𝑟𝑘!

Main results (with reshuffling mobility) Take-away messages : mobility hurts ! Throughput-delay trade-offs are worse than those achievable in static network (in the case of fixed transmission range 𝑅) With power control (transmission range adapted to the content, smaller 𝑝 𝑚 -> larger 𝑅 𝑚 ) we recover exactly the trade-offs of a static network !

Main facts about optimal strategy (with reshuffling mobility) One-hop communications are optimal ! i.e.: wait until you meet a node caching the content you want, and get it Use tx-range(s) such that you compete only with bounded other nodes, and no smaller than this Optimal number of replicas 𝑋 𝑚 are: 𝑋 𝑚 ∝ 𝑝 𝑚 (constant tx range) 𝑋 𝑚 ∝ 𝑝 𝑚 2/3 (variable tx ranges)

The optimal replication (reshuffling mobility) A constrained optimization problem …solved using standard methods

Main results (with random walk mobility) For the random walk mobility model: Multi-hop communications become feasible, provided that flight size F < R (tx-range) As we vary the flight size, we obtain intermediate trade-offs in between the best ones (quasi-static network, F=Θ 1/ 𝑁 ) and the worst ones (fully mobile network, F=Θ 1 ) Using content-adaptable transmission ranges 𝑅 𝑚 we can always recover the best trade-offs of a quasi-static network

Conclusions and future work…with some criticism We derived scaling laws of mobile wireless networks under content-centric (anycast) traffic in one simple scenario Many (too many?) possible extensions: Cache size scaling with 𝑁 Fluid limit (arbitrarily small packets) Communications while moving Dynamic contents with varying popularity But…are these things really interesting?

One possible application… Off-loading of 3G/4G cellular networks by device-to-device opportunistic communications Do we meet enough people with common interests to make it effective? (𝛼<1)‼ Are we patient enough to wait? How about energy cost to keep wireless interfaces up?

Thanks!