Network Codes for Next-Generation Distributed Storage:! Opportunities & Challenges. Kannan Ramchandran
|
|
- Lambert Stewart
- 5 years ago
- Views:
Transcription
1 Network Codes for Next-Generation Distributed Storage:! Opportunities & Challenges Kannan Ramchandran University of California, Berkeley Joint work with: Salim El Rouayheb (UCB), Hao Zhang (UCB), Sameer Pawar (UCB) & Minghua Chen (CUHK)
2 Acknowledgments Many students, former students, colleagues and collaborators have shaped and contributed to this talk. An incomplete list (in alphabetical order) includes: Alex Dimakis (USC) Nima Noorshams (UCB), P. Vijay Kumar & his students N. Shah & K. Rashmi (IISc.), Abhay Parekh (UCB) Changho Suh (UCB) Special thanks to Raymond Yeung & Alfred Ho (CUHK).
3 Data Everywhere Airport Coffee shop Office Home Increasing popularity of data-enabled devices and Hot spots, 4G, Wimax, Wi-fi etc, => always online infrastructure. Truly mobile users Want to access seamlessly and to share data anytime anywhere
4 Danger signs
5 Off To The Cloud Want our files to follow us wherever we go Put them in the cloud Get them from the cloud Data Centers P2P Systems Cloud Applications
6 Server Cloud: is the honeymoon Main Challenge: Server cloud formed of unreliable and untrusted components over? The first and fundamental principle in building robust architecture is to design for failure. Component failures are the norm rather than the exception. Google File System
7 This talk: VoD System with Distributed Caches Distributed Cache Network can be: Set-Top Boxes, home gateways ISP P2P infrastructure nodes: Comcast Idle Internet PC users: Thunder, PPStream, PPLive, Wuala etc Demand distribution Heavy tail Massive video library Cache Network Server Cache Network Goal: Build in reliability and scalability upfront in the architecture
8 Building in reliability and efficiency
9 Three-Tier Architecture: Separation principle Commodity disks Tier I: Server Cloud Tier II: Distributed Cache Network (DCN) Data reliability Serve as life-line Hard guarantees Scalability Adaptivity Soft guarantees Tier III: Users Availability QoS
10 Distributed Cache Network Problem Statement: Minimize server load Subject to: Connectivity constraints Storage constraints Bandwidth constraints Who connects to whom? Choices to be made: What packets to store? Of what movies? Which packets to serve? To which user? Combinatorial explosion: seems to be a hopeless task! Scalability requirement distributed solution
11 Talk Outline I Backdrop & Motivation Scalable & Reliable VoD system II Server Tier Reliability Regenerating codes Uncoded Exact Repair: DRESS codes Min. server load: dist. opt. alg. Adaptivity and growth III Distributed Cache Network
12 A Multi-dimensional Problem Reliability Codes Decoding complexity C rate Traditional Coding Theory Coding for cloud storage
13 Different Codes for Different Folks 2MB File a b n 1 n 2 n 3 Replication a 1MB a b a New node a Download 1 MB (4,2) MDS Erasure Code n 1 n 2 n 3 a b a+b b New node a+b a Download 2MB n 4 b n 4 a+2b tolerates 1 failure tolerates 2 failures Reliability Bandwidth Complexity/ Disk Read
14 Regeneration Codes: Overview Classical approach: download 2 MB 1MB 1MB a But new node (e) is downloading 2 MB to store only 1MB! Q: Possible to download less? a b A: If (e) connects to 2 peers, no But it s possible to download 1.5MB (MSR: Min. Storage Regen. code) b a+b Q: Is this the best possible? a+2b e A: No, by storing a bit more, BW can reduce to 1.2 MB (MBR: Min. BW Regen. Code) When coding is used, creating new packets is not a trivial task. The problem is that to create a new packet we must have access to the entire data object
15 Proof sketch: Information flow graph 1MB a a data collector S a b b c d b c d β β β e data collector α =1MB Dimakis, G., W., W., R β 2 β 1/2MB Total download 1.5MB
16 Proof sketch: reduction to multicasting data collector a a data collector S b c d b c d β β β e data collector data collector data collector data collector Storage Code regeneration is reduced to multicasting on the information flow graph. sufficient iff minimum of the min cuts is larger than file size M. (using Network coding theorem by Yeung et al.) Per-link repair BW Total Repair BW= (M/k)/(1-R) where R= (k-1)/(n-1) Dimakis, G., W., W., R. 07
17 Fundamental Storage/Bandwidth Tradeoff File = 20 MB (40,20) code Storage cost Internet, Wireless MBR Min. Bandwidth Regime Code Storage Bandwidth Savings MDS 1 MB 20 MB MSR 1 MB 2 MB 10x MBR 1.33 MB x Per Failure Savings Optimal Tradeoff can be attained with Random Linear Network Coding MSR Archival applications Min. Storage Regime MDS Repair BW
18 Yearly Savings: Data Centers Google data center in Oregon 800,000 servers Total data= 1.6 million TB Failure rate=4% per year Repair every 2 days 3-replication reliability ( ) Reliability achieved by a [9,6] MBR code Comparison to 3x Replication Yearly Savings Percentage Number Storage 20% 6400 hard disks Bandwidth 20% 6400 TB Why? Same amount of data Replication needs more hard disks More failures higher repair bandwidth
19 Random Linear Network Codes 1mb 1mb a1 a2 b1 b2 1 1 (n,k,d) # nodes contacted for repair b1+b2 # nodes user contacts Total # of nodes a1+b1 a2+b2 a1+2b1 a2+2b a1+b1+2a2+2b2 a1+2b1+3a2+6b2 e1 e2 Is that the end of the story?
20 Exact Regeneration WHY? a b 1. Alphabet Size 2. Systematic form of the data 3. Security (talk by Salim) M (file size) c d e = a We want exact regeneration!
21 Exact repair problem: Non-Multicast a " user File b "... replacement node a, b wants a only a+b user... a+2b user Want everything k choose n users that want all the data n repair nodes that want distinct data A very hard problem in general
22 Tradeoff with Exactness Constraint Storage cost Product-Matrix codes (Rashmi et al.), MBR Focus of this talk (DRESS codes: El Rouayheb & R) Intermediate points not achievable (Shah et al. 10) MSR Asymptotically achievable (Suh & R 10, Cadambe et al. 10): Interference Alignment techniques MDS Repair BW
23 Multiple System Considerations Bandwidth-efficiency comes with a price-tag. 1) Excessive disk I/O delay in repair 2) Complexity: Linear mixing is not free 3) Personal conversation with Jay Wyley (HPL) and Arif Merchant (Google) Can we have codes with min overhead of: 1) Repair bandwidth 2) Disk reads 3) Computational complexity?
24 DRESS codes No data processing for repair: uncoded repair Minimum disk reads for repair Linear encoding/decoding complexity Maximum Distance Separable file MDS Code Fractional Repetition Distributed REplication-based Simple Storage Storage Cloud Surprise: no loss of bandwidth efficiency Bonus: well-suited for security applications System price: repair not as flexible w.r.t. who can help
25 Codes For d=n-1 Min Bandwidth Regime (MBR) (n,k,d)=(4,2,3) 3MB Max file size= 5 MB: cut-set bound file 5MB parity-check MDS Code user Uncoded Repair Minimum reads No processing How about d<n-1? Replacement node Rashmi, Shah, Kumar & R '09
26 Formal DRESS Codes Lines in projective planes intersect in exactly one point Fano Plane Projective plane of order 2 Points Packets Lines Storage nodes
27 Higher Order Projective Planes order m=2 m=3 m=4 Projective planes are guaranteed to exist for any prime power order m.
28 More Codes From Steiner Systems A Steiner system is a collection of points & lines such that: 1. There are points in total 2. Each line contains exactly points We want t=2 3. There is exactly one line that contains any given points Fano plane =S(2, 3, 7) # pts on a line Total # points S(2,3,9)
29 Codes from Steiner Systems Theorem : ( ElRouayheb & R Allerton 10) A Steiner system gives a capacity-achieving DRESS code with parameters using correspondence lines: nodes, points: packets. DRESS Codes? Steiner Systems Projective planes Dual codes? Open problem: DRESS codes beyond Steiner systems????
30 System Design : Formal DRESS Codes in the Server Cloud Stripe 20 MB MDS Code Stripe=20 MB (delay constraint) Edges=packets Outer MDS code rate 0.5 (cache network constraint) DRESS code based on regular graphs Vertices= storage nodes (n,k,d)=(13,5,6) Regular Graph on n=16 Every packet repeated twice vertices and degree d=6 Each node has 6 packets Any 5 nodes have 20 distinct packets (independent linear equations)
31 Talk Roadmap I Background & Motivation II Servers Min. server load: dist. opt. alg. Adaptivity and growth: III Distributed Cache Network
32 Recap: Three-Tier Architecture Tier I: Server Cloud The bottom line: Server Bandwidth Tier II: Cache Nodes Demand distrib. Heavy tail movies Tier III: Users
33 Recap: Distributed Cache Network Goal: Minimize total number of packets served by the server Subject to: Connectivity constraints Storage constraints Bandwidth constraints Who connects to whom? Choices: What packets to store? of which movies? Which packets to serve? to which user? Challenge: Combinatorially explosive optimization problem that needs to be solved in a fully distributed manner
34 Step 1: From Movies to Flavors Use coding to obviate packet scheduling nightmare Manage quantity rather than identity Random Linear Network Codes x 1 x 2 x 1 No Coding x 2 x 2 x 1 x 2 x 2 Coding x 1 +x 2 x 1 +x 2 Who connects to whom? How many What packets to store? How many Which packets to serve and to whom? Combinatorial problem: Hard Convex Optimization Problem: Easy
35 Distributed Cache Network: Architectural Goals Scalability: Match arbitrary demand distribution Robust to node churn/failures Dynamics (fluctuations) Easy to deploy ( plug and play ) Easy to maintain ( disposable ) Server Cloud Self-sustaining : Decentralized: don t bug the server Min. overhead of BW, storage, disk reads, complexity (field-size) Security: avoid creating new packets Desirable to avoid coding at cache layer.
36 What codes do we want at the cache layer? Desired abstraction: Like rateless fountain codes Easy to grow and shrink as needed Maintain quasi-mds property: user gets his data by contacting k (1+ε) nodes But: Distributed Not dependent on large blocklengths Uncoded Caveat: Need to be compatible with the code structure of the server layer Proposed solution: Informal DRESS Codes
37 Informal DRESS Codes M packets Reed- Solomon θ Randomly throw αpackets Outer code of server layer User needs to collect M distinct packets
38 Informal DRESS Code: User Guarantees # nodes user needs to contact: k_min=4, k_avg=5.6, k_max~9 (39,20) outer code 5 packets per node (20,12) outer code 3 packets per node Theorem :(Pawar et al. ISIT`10) Let F be the number of colors observed by a user contacting k nodes then with
39 Decentralized Growing M packets Reed- Solomon θ Minimize the number of knocks on door. Growth simulates rateless fountain code But fully distributed + the cache layer Ex.: α= New cache node
40 Tradeoff Between User and Growth : Intuition M packets Reed- Solomon θ R 0 Cache nodes Code rate R=M/θ Almost no replicas R M Many replicas User User Growth Growth
41 User vs. Growth Tradeoff M packets R=M/θ Reed- Solomon θ Average Analysis: K: # nodes contacted by user E(G) G: # nodes contacted for growth E(K)
42 CACHE nodes USERS
43
44
45
46 Idea: Use Markov Approximation with bounded gap Fully distributed algorithm: soft worst-neighbor choking
47 stay in the max- u.lity configura.on for most of the.me Idea: Use Markov Approximation with bounded gap Fully distributed algorithm: soft worst-neighbor choking sum = 650kbps 100kbps 300kbps Each user- cache node connec8on graph is called a configura8on f System u8lity under f is U f
48
49
50
51 Distributed Load-Balancing
52 Conclusion Massively scalable & reliable distributed content- delivery system Architectural principle: separation of reliability and scalability Role of codes critical: what codes to use? BW, storage, disk I/O, complexity, latency, energy, security DRESS codes: uncoded repair Server Layer: Formal DRESS Codes Deterministic constructions from Proj. Spaces & Steiner system Distributed Cache Layer: Informal DRESS Codes Randomized constructions to enable distributed digital fountain Many open problems in theory, code construction, architecture, system design,
53 Open problems Theory & code construction: Exact- Repair MDS codes for rates above ½? Info- theoretic opt. storage/bw tradeoffs of uncoded repair? DRESS codes: beyond projective planes & Steiner Systems? Fundamental multi- dimensional tradeoffs including system parameters like latency, disk reads, energy, security,? Formal connections to interference alignment in wireless systems? System design and architecture: Incentive mechanisms/economics/game- theory for DCN? Role of caching in wireless systems? Practical systems challenges and testbed deployment?
54 Vision of tomorrow s cloud server? Distributed Cache Network
55 QUESTIONS?
DRESS Code For The Storage Cloud
DRESS Code For The Storage Cloud Distributed Replication-based Exact Simple Storage Salim El Rouayheb EECS Department University of California, Berkeley Joint work with: Sameer Pawar Nima Noorshams Prof.
More informationSCALING UP OF E-MSR CODES BASED DISTRIBUTED STORAGE SYSTEMS WITH FIXED NUMBER OF REDUNDANCY NODES
SCALING UP OF E-MSR CODES BASED DISTRIBUTED STORAGE SYSTEMS WITH FIXED NUMBER OF REDUNDANCY NODES Haotian Zhao, Yinlong Xu and Liping Xiang School of Computer Science and Technology, University of Science
More informationNetwork Coding for Distributed Storage Systems* Presented by Jayant Apte ASPITRG 7/9/13 & 7/11/13
Network Coding for Distributed Storage Systems* Presented by Jayant Apte ASPITRG 7/9/13 & 7/11/13 *Dimakis, A.G.; Godfrey, P.B.; Wu, Y.; Wainwright, M.J.; Ramchandran, K. "Network Coding for Distributed
More informationFractional Repetition Codes for Repair in Distributed Storage Systems
Fractional Repetition Codes for Repair in Distributed Storage Systems Salim El Rouayheb and Kannan Ramchandran Department of Electrical Engineering and Computer Sciences University of California, Bereley
More informationEnabling Node Repair in Any Erasure Code for Distributed Storage
Enabling Node Repair in Any Erasure Code for Distributed Storage K. V. Rashmi, Nihar B. Shah, and P. Vijay Kumar, Fellow, IEEE Abstract Erasure codes are an efficient means of storing data across a network
More informationModern Erasure Codes for Distributed Storage Systems
Modern Erasure Codes for Distributed Storage Systems Storage Developer Conference, SNIA, Bangalore Srinivasan Narayanamurthy Advanced Technology Group, NetApp May 27 th 2016 1 Everything around us is changing!
More informationModern Erasure Codes for Distributed Storage Systems
Modern Erasure Codes for Distributed Storage Systems Srinivasan Narayanamurthy (Srini) NetApp Everything around us is changing! r The Data Deluge r Disk capacities and densities are increasing faster than
More informationExact Optimized-cost Repair in Multi-hop Distributed Storage Networks
Exact Optimized-cost Repair in Multi-hop Distributed Storage Networks Majid Gerami, Ming Xiao Communication Theory Lab, Royal Institute of Technology, KTH, Sweden, E-mail: {gerami, mingx@kthse arxiv:14012774v1
More informationCache-Aided Private Information Retrieval with Partially Known Uncoded Prefetching
Cache-Aided Private Information Retrieval with Partially Known Uncoded Prefetching Yi-Peng Wei Karim Banawan Sennur Ulukus Department of Electrical and Computer Engineering University of Maryland, College
More informationEECS 121: Coding for Digital Communication & Beyond Fall Lecture 22 December 3. Introduction
EECS 121: Coding for Digital Communication & Beyond Fall 2013 Lecture 22 December 3 Lecturer: K. V. Rashmi Scribe: Ajay Shanker Tripathi 22.1 Context Introduction Distributed storage is a deeply relevant
More informationOn Secret Sharing Schemes based on Regenerating Codes
On Secret Sharing Schemes base on Regenerating Coes Masazumi Kurihara (University of Electro-Communications) 0 IEICE General Conference Okayama Japan 0- March 0. (Proc. of 0 IEICE General Conference AS--
More informationInformation-theoretically Secure Regenerating Codes for Distributed Storage
Information-theoretically Secure Regenerating Codes for Distributed Storage Nihar B. Shah, K. V. Rashmi and P. Vijay Kumar Abstract Regenerating codes are a class of codes for distributed storage networks
More informationPRIVATE INFORMATION RETRIEVAL WITH SIDE INFORMATION
PRIVATE INFORMATION RETRIEVAL WITH SIDE INFORMATION SWANAND KADHE, BRENDEN GARCIA, ANOOSHEH HEIDARZADEH, AND ALEX SPRINTSON TEXAS A&M UNIVERSITY SALIM EL ROUAYHEB RUTGERS UNIVERSITY Side Info + Security
More informationRegenerating Codes for Errors and Erasures in Distributed Storage
Regenerating Codes or Errors and Erasures in Distributed Storage K V Rashmi, Nihar B Shah, Kannan Ramchandran, Fellow, IEEE, and P Vijay Kumar, Fellow, IEEE arxiv:0050v csit] 3 May 0 Abstract Regenerating
More informationInterference Alignment in Regenerating Codes for Distributed Storage: Necessity and Code Constructions
2134 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL 58, NO 4, APRIL 2012 Interference Alignment in Regenerating Codes for Distributed Storage: Necessity and Code Constructions Nihar B Shah, K V Rashmi, P
More informationCache-Aided Private Information Retrieval with Unknown and Uncoded Prefetching
Cache-Aided Private Information Retrieval with Unknown and Uncoded Prefetching Yi-Peng Wei Karim Banawan Sennur Ulukus Department of Electrical and Computer Engineering University of Maryland, College
More informationCodes, Bloom Filters, and Overlay Networks. Michael Mitzenmacher
Codes, Bloom Filters, and Overlay Networks Michael Mitzenmacher 1 Today... Erasure codes Digital Fountain Bloom Filters Summary Cache, Compressed Bloom Filters Informed Content Delivery Combining the two
More informationOptimal Exact-Regenerating Codes for Distributed Storage at the MSR and MBR Points via a Product-Matrix Construction
IEEE TRANSACTIONS ON INFORMATION THEORY, VOL 57, NO 8, AUGUST 2011 5227 Optimal Exact-Regenerating Codes for Distributed Storage at the MSR and MBR Points via a Product-Matrix Construction K V Rashmi,
More informationDuality for Simple Multiple Access Networks
Duality for Simple Multiple Access Networks Iwan Duursma Department of Mathematics and Coordinated Science Laboratory U of Illinois at Urbana-Champaign DIMACS Workshop on Network Coding December 15-17,
More informationOn Coding Techniques for Networked Distributed Storage Systems
On Coding Techniques for Networked Distributed Storage Systems Frédérique Oggier frederique@ntu.edu.sg Nanyang Technological University, Singapore First European Training School on Network Coding, Barcelona,
More informationAn Improvement of Quasi-cyclic Minimum Storage Regenerating Codes for Distributed Storage
An Improvement of Quasi-cyclic Minimum Storage Regenerating Codes for Distributed Storage Chenhui LI*, Songtao LIANG* * Shanghai Key Laboratory of Intelligent Information Processing, Fudan University,
More informationA survey on regenerating codes
International Journal of Scientific and Research Publications, Volume 4, Issue 11, November 2014 1 A survey on regenerating codes V. Anto Vins *, S.Umamageswari **, P.Saranya ** * P.G Scholar, Department
More informationAn Optimized Video-on-Demand System: Theory, Design and Implementation
An Optimized Video-on-Demand System: Theory, Design and Implementation Hao Zhang Electrical Engineering and Computer Sciences University of California at Berkeley Technical Report No. UCB/EECS-202-229
More informationBASIC Regenerating Code: Binary Addition and Shift for Exact Repair
BASIC Regenerating Code: Binary Addition and Shift for Exact Repair Hanxu Hou, Kenneth W. Shum, Minghua Chen and Hui Li Shenzhen Eng. Lab of Converged Networks Technology, Shenzhen Key Lab of Cloud Computing
More informationCoding Techniques for Distributed Storage Systems
Coding Techniques for Distributed Storage Systems Jung-Hyun Kim, {jh.kim06}@yonsei.ac.kr, Hong-Yeop Song Yonsei Univ. Seoul, KORE 3 rd CITW, Oct. 25. 1, 2013 Contents Introduction Codes for Distributed
More informationBATS: Achieving the Capacity of Networks with Packet Loss
BATS: Achieving the Capacity of Networks with Packet Loss Raymond W. Yeung Institute of Network Coding The Chinese University of Hong Kong Joint work with Shenghao Yang (IIIS, Tsinghua U) R.W. Yeung (INC@CUHK)
More informationGraph based codes for distributed storage systems
/23 Graph based codes for distributed storage systems July 2, 25 Christine Kelley University of Nebraska-Lincoln Joint work with Allison Beemer and Carolyn Mayer Combinatorics and Computer Algebra, COCOA
More informationCooperative Pipelined Regeneration in Distributed Storage Systems
Cooperative ipelined Regeneration in Distributed Storage Systems Jun Li, in Wang School of Computer Science Fudan University, China Baochun Li Department of Electrical and Computer Engineering University
More informationComputing over Multiple-Access Channels with Connections to Wireless Network Coding
ISIT 06: Computing over MACS 1 / 20 Computing over Multiple-Access Channels with Connections to Wireless Network Coding Bobak Nazer and Michael Gastpar Wireless Foundations Center Department of Electrical
More informationMinimization of Storage Cost in Distributed Storage Systems with Repair Consideration
Minimization of Storage Cost in Distributed Storage Systems with Repair Consideration Quan Yu Department of Electronic Engineering City University of Hong Kong Email: quanyu2@student.cityu.edu.hk Kenneth
More informationAn Information-Theoretic Perspective of Consistent Distributed Storage
An Information-Theoretic Perspective of Consistent Distributed Storage Viveck R. Cadambe Pennsylvania State University Joint with Prof. Zhiying Wang (UCI) Prof. Nancy Lynch (MIT), Prof. Muriel Medard (MIT)
More informationPrivate Information Retrieval from MDS Coded Data in Distributed Storage Systems
Private Information Retrieval from MDS Coded Data in Distributed Storage Systems Razan Tajeddine Salim El Rouayheb ECE Department IIT Chicago Emails: rtajeddi@hawiitedu salim@iitedu arxiv:160201458v1 [csit]
More informationPerformance Models of Access Latency in Cloud Storage Systems
Performance Models of Access Latency in Cloud Storage Systems Qiqi Shuai Email: qqshuai@eee.hku.hk Victor O.K. Li, Fellow, IEEE Email: vli@eee.hku.hk Yixuan Zhu Email: yxzhu@eee.hku.hk Abstract Access
More informationCodes for Modern Applications
Codes for Modern Applications Current Research Topics in the Code and Signal Design Group P. Vijay Kumar Indian Institute of Science, Bangalore, India Nov. 16, 2018 1/28 Codes for Modern Applications:
More informationNash Equilibrium Load Balancing
Nash Equilibrium Load Balancing Computer Science Department Collaborators: A. Kothari, C. Toth, Y. Zhou Load Balancing A set of m servers or machines. A set of n clients or jobs. Each job can be run only
More informationData Integrity Protection scheme for Minimum Storage Regenerating Codes in Multiple Cloud Environment
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Data Integrity Protection scheme for Minimum Storage Regenerating Codes in Multiple Cloud Environment Saravana
More informationPrivacy, Security, and Repair in Distributed Storage Systems. Siddhartha Kumar
Privacy, Security, and Repair in Distributed Storage Systems Siddhartha Kumar Thesis for the Degree of Philosophiae Doctor (PhD) University of Bergen, Norway 2018 Privacy, Security, and Repair in Distributed
More informationStream Sessions: Stochastic Analysis
Stream Sessions: Stochastic Analysis Hongwei Zhang http://www.cs.wayne.edu/~hzhang Acknowledgement: this lecture is partially based on the slides of Dr. D. Manjunath and Dr. Kumar Outline Loose bounds
More informationCSE 4/60373: Multimedia Systems
CSE 4/60373: Multimedia Systems Outline for today 32: Y.-F. Chen, Y. Huang, R. Jana, H. Jiang, M. Rabinovich, J. Rahe, B. Wei, and Z. Xiao. Towards Capacity and Profit Optimization of Video-on-Demand Services
More informationYour Data is in the Cloud: Who Exactly is Looking After It?
Your Data is in the Cloud: Who Exactly is Looking After It? P Vijay Kumar Dept of Electrical Communication Engineering Indian Institute of Science IISc Open Day March 4, 2017 1/33 Your Data is in the Cloud:
More informationA New Combinatorial Design of Coded Distributed Computing
A New Combinatorial Design of Coded Distributed Computing Nicholas Woolsey, Rong-Rong Chen, and Mingyue Ji Department of Electrical and Computer Engineering, University of Utah Salt Lake City, UT, USA
More informationRandomized User-Centric Clustering for Cloud Radio Access Network with PHY Caching
Randomized User-Centric Clustering for Cloud Radio Access Network with PHY Caching An Liu, Vincent LAU and Wei Han the Hong Kong University of Science and Technology Background 2 Cloud Radio Access Networks
More informationLecture 19. Lecturer: Aleksander Mądry Scribes: Chidambaram Annamalai and Carsten Moldenhauer
CS-621 Theory Gems November 21, 2012 Lecture 19 Lecturer: Aleksander Mądry Scribes: Chidambaram Annamalai and Carsten Moldenhauer 1 Introduction We continue our exploration of streaming algorithms. First,
More informationCache and Forward Architecture
Cache and Forward Architecture Shweta Jain Research Associate Motivation Conversation between computers connected by wires Wired Network Large content retrieval using wireless and mobile devices Wireless
More informationAll About Erasure Codes: - Reed-Solomon Coding - LDPC Coding. James S. Plank. ICL - August 20, 2004
All About Erasure Codes: - Reed-Solomon Coding - LDPC Coding James S. Plank Logistical Computing and Internetworking Laboratory Department of Computer Science University of Tennessee ICL - August 2, 24
More informationJoint Server Selection and Routing for Geo-Replicated Services
Joint Server Selection and Routing for Geo-Replicated Services Srinivas Narayana Joe Wenjie Jiang, Jennifer Rexford and Mung Chiang Princeton University 1 Large-scale online services Search, shopping,
More informationErasure Coding in Object Stores: Challenges and Opportunities
Erasure Coding in Object Stores: Challenges and Opportunities Lewis Tseng Boston College July 2018, PODC Acknowledgements Nancy Lynch Muriel Medard Kishori Konwar Prakash Narayana Moorthy Viveck R. Cadambe
More informationDesign and Performance Analysis of a DRAM-based Statistics Counter Array Architecture
Design and Performance Analysis of a DRAM-based Statistics Counter Array Architecture Chuck Zhao 1 Hao Wang 2 Bill Lin 2 Jim Xu 1 1 Georgia Institute of Technology 2 University of California, San Diego
More informationEfficient Content Delivery and Low Complexity Codes. Amin Shokrollahi
Efficient Content Delivery and Low Complexity Codes Amin Shokrollahi Content Goals and Problems TCP/IP, Unicast, and Multicast Solutions based on codes Applications Goal Want to transport data from a transmitter
More informationCutting the Cord: A Robust Wireless Facilities Network for Data Centers
Cutting the Cord: A Robust Wireless Facilities Network for Data Centers Yibo Zhu, Xia Zhou, Zengbin Zhang, Lin Zhou, Amin Vahdat, Ben Y. Zhao and Haitao Zheng U.C. Santa Barbara, Dartmouth College, U.C.
More informationWhen Locally Repairable Codes Meet Regenerating Codes What If Some Helpers Are Unavailable
When Locally Repairable Codes Meet Regenerating Codes What If Some Helpers Are Unavailable Imad Ahmad, Chih-Chun Wang; {ahmadi,chihw}@purdue.edu School of Electrical and Computer Engineering, Purdue University,
More informationDeveloping MapReduce Programs
Cloud Computing Developing MapReduce Programs Dell Zhang Birkbeck, University of London 2017/18 MapReduce Algorithm Design MapReduce: Recap Programmers must specify two functions: map (k, v) * Takes
More informationDistributed Video Systems Chapter 3 Storage Technologies
Distributed Video Systems Chapter 3 Storage Technologies Jack Yiu-bun Lee Department of Information Engineering The Chinese University of Hong Kong Contents 3.1 Introduction 3.2 Magnetic Disks 3.3 Video
More informationBCStore: Bandwidth-Efficient In-memory KV-Store with Batch Coding. Shenglong Li, Quanlu Zhang, Zhi Yang and Yafei Dai Peking University
BCStore: Bandwidth-Efficient In-memory KV-Store with Batch Coding Shenglong Li, Quanlu Zhang, Zhi Yang and Yafei Dai Peking University Outline Introduction and Motivation Our Design System and Implementation
More informationPrivate Information Retrieval with Side Information: the Single Server Case
Private Information Retrieval with Side Information: the Single Server Case Swanand adhe, Brenden Garcia, Anoosheh Heidarzadeh, Salim El Rouayheb, and Alex Sprintson Abstract We study the problem of Private
More informationThe Encoding Complexity of Network Coding
The Encoding Complexity of Network Coding Michael Langberg Alexander Sprintson Jehoshua Bruck California Institute of Technology Email: mikel,spalex,bruck @caltech.edu Abstract In the multicast network
More informationSummary of Raptor Codes
Summary of Raptor Codes Tracey Ho October 29, 2003 1 Introduction This summary gives an overview of Raptor Codes, the latest class of codes proposed for reliable multicast in the Digital Fountain model.
More informationSome problems in ad hoc wireless networking. Balaji Prabhakar
Some problems in ad hoc wireless networking Balaji Prabhakar Background Example scenarios for ad hoc packet networks - sensor networks (many nodes, low data rates) - wireless LANs (fewer nodes, high data
More informationIntroduction to Algorithms / Algorithms I Lecturer: Michael Dinitz Topic: Algorithms and Game Theory Date: 12/3/15
600.363 Introduction to Algorithms / 600.463 Algorithms I Lecturer: Michael Dinitz Topic: Algorithms and Game Theory Date: 12/3/15 25.1 Introduction Today we re going to spend some time discussing game
More informationA Scalable Content- Addressable Network
A Scalable Content- Addressable Network In Proceedings of ACM SIGCOMM 2001 S. Ratnasamy, P. Francis, M. Handley, R. Karp, S. Shenker Presented by L.G. Alex Sung 9th March 2005 for CS856 1 Outline CAN basics
More informationArchitekturen für die Cloud
Architekturen für die Cloud Eberhard Wolff Architecture & Technology Manager adesso AG 08.06.11 What is Cloud? National Institute for Standards and Technology (NIST) Definition On-demand self-service >
More informationCoding for Improved Throughput Performance in Network Switches
Coding for Improved Throughput Performance in Network Switches Rami Cohen, Graduate Student Member, IEEE, and Yuval Cassuto, Senior Member, IEEE 1 arxiv:1605.04510v1 [cs.ni] 15 May 2016 Abstract Network
More informationA Unified Coding Framework for Distributed Computing with Straggling Servers
A Unified Coding Framewor for Distributed Computing with Straggling Servers Songze Li, Mohammad Ali Maddah-Ali, and A. Salman Avestimehr Department of Electrical Engineering, University of Southern California,
More informationScalable Video Transport over Wireless IP Networks. Dr. Dapeng Wu University of Florida Department of Electrical and Computer Engineering
Scalable Video Transport over Wireless IP Networks Dr. Dapeng Wu University of Florida Department of Electrical and Computer Engineering Bandwidth Fluctuations Access SW Domain B Domain A Source Access
More informationMohammad Hossein Manshaei 1393
Mohammad Hossein Manshaei manshaei@gmail.com 1393 Voice and Video over IP Slides derived from those available on the Web site of the book Computer Networking, by Kurose and Ross, PEARSON 2 Multimedia networking:
More informationSingle Video Performance Analysis for Video-on-Demand Systems
Single Video Performance Analysis for Video-on-Demand Systems James Y. Yang Department of ECE Coordinated Science Laboratory University of Illinois at Urbana-Champaign 138 W Main Street Urbana, IL 6181
More informationNetworks: A Deterministic Approach with Constant Overhead
Trace-Routing in 3D Wireless Sensor Networks: A Deterministic Approach with Constant Overhead Su Xia, Hongyi Wu, and Miao Jin! School of Computing and Informatics! University of Louisiana at Lafayette
More informationLinear Block Codes. Allen B. MacKenzie Notes for February 4, 9, & 11, Some Definitions
Linear Block Codes Allen B. MacKenzie Notes for February 4, 9, & 11, 2015 This handout covers our in-class study of Chapter 3 of your textbook. We ll introduce some notation and then discuss the generator
More informationNear Optimal Broadcast with Network Coding in Large Sensor Networks
in Large Sensor Networks Cédric Adjih, Song Yean Cho, Philippe Jacquet INRIA/École Polytechnique - Hipercom Team 1 st Intl. Workshop on Information Theory for Sensor Networks (WITS 07) - Santa Fe - USA
More informationCutting the Cord: A Robust Wireless Facilities Network for Data Centers
Cutting the Cord: A Robust Wireless Facilities Network for Data Centers Yibo Zhu, Xia Zhou, Zengbin Zhang, Lin Zhou, Amin Vahdat, Ben Y. Zhao and Haitao Zheng U.C. Santa Barbara, Dartmouth College, U.C.
More informationBackbone Modeling for Carrying Local Content and Over-the-Top Traffic
White Paper Backbone Modeling for Carrying Local Content and Over-the-Top Traffic Decision-Making Criteria Using Cisco MATE Collector and Cisco MATE Design and Their Impact on Backbone Design What You
More informationNovel Decentralized Coded Caching through Coded Prefetching
ovel Decentralized Coded Caching through Coded Prefetching Yi-Peng Wei Sennur Ulukus Department of Electrical and Computer Engineering University of Maryland College Park, MD 2072 ypwei@umd.edu ulukus@umd.edu
More informationResume Maintaining System for Referral Using Cloud Computing
Resume Maintaining System for Referral Using Cloud Computing Pranjali A. Pali 1, N.D. Kale 2 P.G. Student, Department of Computer Engineering, TSSM S PVPIT, Bavdhan, Pune, Maharashtra India 1 Associate
More informationInterdomain Routing Design for MobilityFirst
Interdomain Routing Design for MobilityFirst October 6, 2011 Z. Morley Mao, University of Michigan In collaboration with Mike Reiter s group 1 Interdomain routing design requirements Mobility support Network
More informationMoB: A Mobile Bazaar for Wide Area Wireless Services. R.Chakravorty, S.Agarwal, S.Banerjee and I.Pratt mobicom 2005
MoB: A Mobile Bazaar for Wide Area Wireless Services R.Chakravorty, S.Agarwal, S.Banerjee and I.Pratt mobicom 2005 What is MoB? It is an infrastructure for collaborative wide-area wireless data services.
More informationToday s Papers. Array Reliability. RAID Basics (Two optional papers) EECS 262a Advanced Topics in Computer Systems Lecture 3
EECS 262a Advanced Topics in Computer Systems Lecture 3 Filesystems (Con t) September 10 th, 2012 John Kubiatowicz and Anthony D. Joseph Electrical Engineering and Computer Sciences University of California,
More informationCS246: Mining Massive Datasets Jure Leskovec, Stanford University
CS246: Mining Massive Datasets Jure Leskovec, Stanford University http://cs246.stanford.edu [Kumar et al. 99] 2/13/2013 Jure Leskovec, Stanford CS246: Mining Massive Datasets, http://cs246.stanford.edu
More informationChallenges in the Wide-area. Tapestry: Decentralized Routing and Location. Global Computation Model. Cluster-based Applications
Challenges in the Wide-area Tapestry: Decentralized Routing and Location System Seminar S 0 Ben Y. Zhao CS Division, U. C. Berkeley Trends: Exponential growth in CPU, b/w, storage Network expanding in
More informationSymmetric Private Information Retrieval For MDS Coded Distributed Storage
Symmetric Private Information Retrieval For DS Coded Distributed Storage Qiwen Wang, and ikael Skoglund School of Electrical Engineering, KTH Royal Institute of Technology Email: {qiwenw, skoglund}@kthse
More informationPRESENTED BY SARAH KWAN NETWORK CODING
PRESENTED BY SARAH KWAN NETWORK CODING NETWORK CODING PRESENTATION OUTLINE What is Network Coding? Motivation and Approach Network Coding with Lossless Networks Challenges in Developing Coding Algorithms
More informationOn the Scalability of Hierarchical Ad Hoc Wireless Networks
On the Scalability of Hierarchical Ad Hoc Wireless Networks Suli Zhao and Dipankar Raychaudhuri Fall 2006 IAB 11/15/2006 Outline Motivation Ad hoc wireless network architecture Three-tier hierarchical
More informationDiskReduce: Making Room for More Data on DISCs. Wittawat Tantisiriroj
DiskReduce: Making Room for More Data on DISCs Wittawat Tantisiriroj Lin Xiao, Bin Fan, and Garth Gibson PARALLEL DATA LABORATORY Carnegie Mellon University GFS/HDFS Triplication GFS & HDFS triplicate
More informationApproximation Algorithms
Approximation Algorithms Group Members: 1. Geng Xue (A0095628R) 2. Cai Jingli (A0095623B) 3. Xing Zhe (A0095644W) 4. Zhu Xiaolu (A0109657W) 5. Wang Zixiao (A0095670X) 6. Jiao Qing (A0095637R) 7. Zhang
More informationResearch on Transmission Based on Collaboration Coding in WSNs
Research on Transmission Based on Collaboration Coding in WSNs LV Xiao-xing, ZHANG Bai-hai School of Automation Beijing Institute of Technology Beijing 8, China lvxx@mail.btvu.org Journal of Digital Information
More informationBusiness Benefits of Policy Based Data De-Duplication Data Footprint Reduction with Quality of Service (QoS) for Data Protection
Data Footprint Reduction with Quality of Service (QoS) for Data Protection By Greg Schulz Founder and Senior Analyst, the StorageIO Group Author The Green and Virtual Data Center (Auerbach) October 28th,
More informationApplication Provisioning in Fog Computingenabled Internet-of-Things: A Network Perspective
Application Provisioning in Fog Computingenabled Internet-of-Things: A Network Perspective Ruozhou Yu, Guoliang Xue, and Xiang Zhang Arizona State University Outlines Background and Motivation System Modeling
More informationMatrix algorithms: fast, stable, communication-optimizing random?!
Matrix algorithms: fast, stable, communication-optimizing random?! Ioana Dumitriu Department of Mathematics University of Washington (Seattle) Joint work with Grey Ballard, James Demmel, Olga Holtz, Robert
More informationA Survey on Erasure Coding Techniques for Cloud Storage System
A Survey on Erasure Coding Techniques for Cloud Storage System Kiran P. Pawar #1, R. M. Jogdand *2 #1 M.Tech, Department of Computer Science, Gogte Institute of technology, Udyambag Belagavi, Karnataka,
More informationDesign and Performance Analysis of a DRAM-based Statistics Counter Array Architecture
Design and Performance Analysis of a DRAM-based Statistics Counter Array Architecture Chuck Zhao 1 Hao Wang 2 Bill Lin 2 1 1 Georgia Tech 2 UCSD October 2nd, 2008 Broader High-Level Question What are the
More informationCodes for distributed storage from 3 regular graphs
Codes for distributed storage from 3 regular graphs Shuhong Gao, Fiona Knoll, Felice Manganiello, and Gretchen Matthews Clemson University arxiv:1610.00043v1 [cs.it] 30 Sep 2016 October 4, 2016 Abstract
More informationOn Data Parallelism of Erasure Coding in Distributed Storage Systems
On Data Parallelism of Erasure Coding in Distributed Storage Systems Jun Li, Baochun Li Department of Electrical and Computer Engineering, University of Toronto, Canada {junli, bli}@ece.toronto.edu Abstract
More informationRandomized Algorithms for Network Security and Peer-to-Peer Systems
Randomized Algorithms for Network Security and Peer-to-Peer Systems Micah Adler University of Massachusetts, Amherst Talk Outline Probabilistic Packet Marking for IP Traceback Network Security Appeared
More informationApproximation Algorithms
Chapter 8 Approximation Algorithms Algorithm Theory WS 2016/17 Fabian Kuhn Approximation Algorithms Optimization appears everywhere in computer science We have seen many examples, e.g.: scheduling jobs
More informationBenefits of Coded Placement for Networks with Heterogeneous Cache Sizes
Benefits of Coded Placement for Networks with Heterogeneous Cache Sizes Abdelrahman M. Ibrahim, Ahmed A. Zewail, and Aylin Yener ireless Communications and Networking Laboratory CAN Electrical Engineering
More informationCoding theory for scalable media delivery
1 Coding theory for scalable media delivery Michael Luby RaptorQ is a product of Qualcomm Technologies, Inc. Application layer erasure coding complements traditional error coding Forward Error Correction
More informationFrom Routing to Traffic Engineering
1 From Routing to Traffic Engineering Robert Soulé Advanced Networking Fall 2016 2 In the beginning B Goal: pair-wise connectivity (get packets from A to B) Approach: configure static rules in routers
More information/ Approximation Algorithms Lecturer: Michael Dinitz Topic: Linear Programming Date: 2/24/15 Scribe: Runze Tang
600.469 / 600.669 Approximation Algorithms Lecturer: Michael Dinitz Topic: Linear Programming Date: 2/24/15 Scribe: Runze Tang 9.1 Linear Programming Suppose we are trying to approximate a minimization
More informationCoded Caching for Hierarchical Networks with a Different Number of Layers
Coded Caching for Hierarchical Networks with a Different Number of Layers Makoto Takita, Masanori Hirotomo, Masakatu Morii Kobe University, Saga University November 20, 2017 ASON 17@Aomori Outline 1 1.
More informationDiversity Coloring for Distributed Storage in Mobile Networks
Diversity Coloring for Distributed Storage in Mobile Networks Anxiao (Andrew) Jiang and Jehoshua Bruck California Institute of Technology Abstract: Storing multiple copies of files is crucial for ensuring
More informationLocally repairable codes
Locally repairable codes for large-scale storage Presented by Anwitaman Datta Nanyang Technological University, Singapore Joint work with Frédérique Oggier & Lluís Pàmies i Juárez @ IISc 2012 & NetApp
More information