• Participants: Nha, Dileep, etc
• When: Every week, Thu 3pm (or Fri 2pm)
• Where: TBA
• Topics: Research topics of participants
• Presentor's role:
  • Update agenda table & upload materials in the webpage before meeting
  • Upload updated material if necessary
  • Make sure to do TODO list indicated in the meeting note
• Participants' role:
  • Join the After-Meeting-Discussion (AMD)
    • Provide feedback in AMD
• Kaist's wiki
  • http://itcknow4.kaist.ac.kr:3000/projects/cloudlet/wiki
  • id : cloudlet
  • pw : guest123

• difference from prior work
  • they ignored unreachable for computing cloak region, boosting up success rate
  • expand cloaked area until it meets the min cloaked region area

• why a request can fail ?
  • if the following two gaols are not met until time expires
    • min area constraints cannot be achieved even after expansion (our idea)
    • max-clique size is smaller than max k, even after we repeatedly remove highest-k user (prior work’s idea)

• limitations of conference/thesis work
  • movement simulator (Thosmas Brinkhoff) is not realistic
  • BUT WHY?
    • one-lane two-way movement
    • no traffic signal

• Improvements
  • use Paramics for
    • better realistic car movements
    • movement model?
  • query submission model using exponential distribution
  • utility metrics:
    • current : lifetime
    • new: max acceptable area

• [MH] Todo list
  • Check the movement model of Thosmas Brinkhoff's simulator
  • Re-think the privacy metric of vehicle-based privacy
  • Check pedestrian mode of Paramics

• Attacks on AMI
  • False data injection attack
  • DoS attack
  • Eavesdropping attack
    • Privacy threat
    • Vacancy check leading to intrusive threat
    • TODO: type of information handled by AMI (stored, communicated)
      • Pricing Information, Consumption Readings, Person's Living patterns, Health Insurance can determine the usage of medical devices, landlord can tell no. of peoples staying and their time or parties et.
      • Information Handled by AMI
  • Impersonation attack (MIM)
  • Replay attack
  • TODO: need a detailed discussion of related threats to user & system caused by these attacks
    • These attacks results in sending incorrect data to control centers. Incorrect decisions made by control center may lead to: Inefficient Dispatch / Power Quality Degradation, Unnecessary Load Shedding, False Alarm / Erroneous Fault Detection, System Instability, Financial losses. And if there are many attacks then it may cause to destroy the appliances.

• Dileep's TODO
  • Read and present Set-K-Cover paper by next meeting
  • Nha also read the paper

• no traffic lights, no lane concept
• car speed is unrealistic
  • unlimited lanes
  • car speed linearly increase until hit the max
• speed unit is weird
• starting/ending nodes cannot be controlled

• Literature Survey Guideline
  • Please gather list of papers about the following area
    • distributed computing
    • parallel computing
    • distributed programming model
    • parallel programming model
    • distributed sensing
    • collaborative sensing framework
    • etc
  • The research questions we are interested in are…
    • how can we model a task (sensing task, computation task) that is runnable in a distributed fashion?
    • how can we represent such a task ?
    • what is the programming model for distributed computing/sensing?
    • given a computation task, how can we divide the task into subtasks?
    • how can we assign each subtask to helpers with what criteria with what goal?
    • how can we perform a sensing/computation task in a distributed fashion with what goals?
    • how can we handle exceptions?
    • in what framework a subtask can be delegated to other node?
    • etc
  • For each paper, prepare the following information
    • authors and affiliations
    • title
    • journal/conference and year/vol/issue
    • short description in one or two sentences
  • Known keywords
    • "MapReduce & Hadoop" for distributed computing framework
    • "Medusa" for crowd sensing framework

• MapReduce: Simplified Data Processing on Large Clusters PDF
  • Jeffrey Dean and Sanjay Ghemawat
  • Google Inc
  • ACM 51 (1) (2008) 107–113.
  • The authors proposes a simple and powerful interface that enables automatic parallelization and distribution of large-scale computations, combined with an implementation of this interface that achieves high performance on large clusters of commodity PCs.

• MapReduce optimization algorithm based on machine learning in heterogeneous cloud environment PDF
  • LIN Wen-hui, LEI Zhen-ming, LIU Jun,YANG Jie, LIU Fang, HE Gang, WANG Qin
  • Beijing University of Posts and Telecommunications, China
  • The Journal of China Universities of Posts and Telecommunications (ELSEVIER), Dec 2013
  • The authors develop a machine learning module to obtain nodes’ performance values and optimize the reduce task assignment algorithm & the speculative execution mechanism to improve the performance of heterogeneous clusters.

• A genetic algorithm for task scheduling on heterogeneous computing systems using multiple priority queues PDF
  • Yuming Xu, Kenli Li, Jingtong Hu, Keqin Li
  • College of Information Science and Engineering, Hunan University, China
  • Information Sciences Volume 270, 20 June 2014, Pages 255–287
  • In this paper, a task scheduling scheme on heterogeneous computing systems using a multiple priority queues genetic algorithm (MPQGA) is proposed.

• Performance Issues of Heterogeneous Hadoop Clusters in Cloud Computing PDF
  • B.Thirumala Rao, N.V.Sridevi, V.Krishna Reddy, L.S.S.Reddy
  • Department of Computer Science and Engineering, Lakireddy Bali Reddy College of Engineering,Mylavaram
  • Global Journal of Computer Science and Technology, May 2011, Volume XI Issue VIII
  • The overview of Hadoop and some issues that affect the performance of hadoop in heterogeneous clusters in cloud environments. The paper proposed some guidelines on how to overcome these issues to improve the performance of hadoop.

• SHadoop: Improving MapReduce performance by optimizing job execution mechanism in Hadoop clusters PDF
  • Rong Gu, Xiaoliang Yang, Jinshuang Yan, Yuanhao Sun, Bing Wang, Chunfeng Yuan, Yihua Huang
  • Nanjing University, Intel Asia-Pacific Research and Development Ltd, China
  • Journal of Parallel and Distributed Computing, Vol 74, Issue 3, March 2014, Pages 2166–2179
  • The paper analyzed and identified two critical limitations of MapReduce execution mechanism, achieved first optimization by implementing new job setup/cleanup task, and replaced heartbeat with an instant messaging mechanism to speedup task scheduling.

• Two new fast heuristics for mapping parallel applications on cloud computing PDF
  • I. De Falco, U. Scafuri, E. Tarantino
  • Institute of High Performance Computing and Networking, National Research Council of Italy
  • Future Generation Computer Systems Volume 37, July 2014, Pages 1–13
  • In this paper two new heuristics, named Min–min-C and Max–min-C, are proposed able to provide near-optimal solutions to the mapping of parallel applications, modeled as Task Interaction Graphs, on computational clouds.

• Time-optimized contextual information forwarding in mobile sensor networks PDF
  • Christos Anagnostopoulos
  • School of Computing Science, University of Glasgowm, UK
  • Journal of Parallel and Distributed Computing Volume 74, Issue 5, May 2014, Pages 2317–2332
  • The paper studies on the scheduling of Context Forwarding (CF) in Mobile Sensor Networks (MSNs), formulates the CF scheduling problem as an optimal stopping time problem, proposes an optimal CF policy over MSN, and proposes policy exhibits efficient CF in MSN.

• Method and system for distributed computation having sub-task processing and sub-solution redistribution PDFhttp://www.google.co.uk/patents/US8321870
  • Daniel John Messier, Joseph James Salvo, John William Carbone, Charles Burton Theurer, Li Zhang
  • General Electric Company
  • Patent US8321870 B2, 2012
  • A system for processing a computational task is presented.

• Multi-heuristic dynamic task allocation using genetic algorithms in a heterogeneous distributed system PDF
  • Andrew J. Page, Thomas M. Keane, Thomas J. Naughton
  • Wellcome Trust Sanger Institute in Cambridge
  • Journal of Parallel and Distributed Computing Volume 70, Issue 7, July 2010, Pages 758–766
  • They present a multi-heuristic evolutionary task allocation algorithm to dynamically map tasks to processors in a heterogeneous distributed system.

• An effective and robust two-phase resource allocation scheme for interdependent tasks in mobile ad hoc computational Grids PDF
  • Sayed Chhattan Shah, Qurat-Ul-Ain Nizamani, Sajjad Hussain Chauhdary, Myong-Soon Park
  • Electronics and Telecommunications Research Institute in Daejeon, South Korea
  • Journal of Parallel and Distributed Computing Volume 72, Issue 12, December 2012, Pages 1664–1679
  • They propose a two-phase resource allocation scheme to reduce communication cost between dependent tasks

• An effective iterated greedy algorithm for reliability-oriented task allocation in distributed computing systems PDF
  • Qinma Kang, Hong He, Jun Wei
  • Shandong University, China
  • Journal of Parallel and Distributed Computing Volume 73, Issue 8, August 2013, Pages 1106–1115
  • This paper investigates the problem of allocating parallel application tasks to processors in heterogeneous distributed computing systems with the goal of maximizing the system reliability.

• Ganga: A tool for computational-task management and easy access to Grid resources PDF
  • J.T. Mościcki, F. Brochu, J. Ebke,…
  • CERN, Geneva, Switzerland; University of Cambridge, United Kingdom
  • Computer Physics Communications Volume 180, Issue 11, November 2009, Pages 2303–2316
  • Present the computational task-management tool Ganga, which allows for the specification, submission, bookkeeping and post-processing of computational tasks on a wide set of distributed resources.

• Dryad: distributed data-parallel programs from sequential building blocks PDF
  • Michael Isard, Mihai Budiu, Yuan Yu, Andrew Birrell, Dennis Fetterly
  • Microsoft Research, Silicon Valley
  • ACM SIGOPS Operating Systems Review - EuroSys'07 Conference Proceedings Volume 41 Issue 3, June 2007 Pages 59-72
  • Dryad is a popular programming model for implementing parallel and distributed programs that can scale up capability of processing from a very small cluster to a large cluster.

• Optimal task partition and distribution in grid service system with common cause failures PDF
  • Yuan-Shun Dai, Gregory Levitin, Xiaolong Wang
  • Department of Computer and Information Science, Purdue University, USA
  • Future Generation Computer Systems Volume 23, Issue 2, February 2007, Pages 209–218
  • This paper solved an important optimization problem for grid service. The problem was to maximize the expected profit by partitioning the service task into subtasks and by distributing them among the available resources. A genetic algorithm was presented to solve this type of optimization problem.

• Dynamic tuning of the workload partition factor and the resource utilization in data-intensive applications PDF
  • Claudia Rosas, Anna Sikora, Josep Jorba, Andreu Moreno, Antonio Espinosa, Eduardo César
  • Universitat Autònoma de Barcelona, Spain
  • Future Generation Computer Systems Volume 37, July 2014, Pages 162–177
  • This work proposes a methodology to dynamically improve the performance of data-intensive applications based on: (i) adapting the size and the number of data partitions to reduce the overall execution time; and (ii) adapting the number of processing nodes to achieve an efficient execution.

• An energy-efficient process clustering assignment algorithm for distributed system PDF
  • Anan Niyom, Peraphon Sophatsathit , Chidchanok Lursinsap
  • Chulalongkorn University, Thailand
  • Simulation Modelling Practice and Theory Volume 40, January 2014, Pages 95–111
  • A new set of task scheduling and assignment algorithms focusing on minimizing energy consumption, named Energy-Efficient Process Clustering Assignment algorithm or EPC algorithm, was proposed.

• Parallel partitioning for distributed systems using sequential assignment PDF
  • Simon Spacey, Wayne Luk, Daniel Kuhn, Paul H.J. Kelly
  • Department of Computing, Imperial College London, UK
  • Journal of Parallel and Distributed Computing Volume 73, Issue 2, February 2013, Pages 207–219
  • This paper introduces a method to combine the advantages of both task parallelism and fine-grained co-design specialisation to achieve faster execution times than either method alone on distributed heterogeneous architectures.

• Two-phase grouping-based resource management for big data processing in mobile cloud computing PDF
  • JiSu Park, Hyongsoon Kim, Young-Sik Jeong and Eunyoung Lee
  • Korea University, Dongduk Womens University, Korea
  • INTERNATIONAL JOURNAL OF COMMUNICATION SYSTEMS 2013
  • This paper proposes a grouping technique based on the utilization and movement rates to cope the movement problem and the utilization problem of mobile devices in cloud computing.

• Programming support and scheduling for communicating parallel tasks PDF
  • Jörg Dümmler, Thomas Rauber, Gudula Rünger
  • Chemnitz University of Technology, Germany
  • Journal of Parallel and Distributed Computing Volume 73, Issue 2, February 2013, Pages 220–234
  • This paper describes the parallel programming model of communicating M-tasks (CM-tasks), discusses the scheduling of CM-task programs and presents an appropriate algorithm.

• Distributed and Cloud Computing: From Parallel Processing to the Internet of things PDF
  • Kai Hwang, Jack Dongarra, Geoffrey C. Fox
  • Univ. of Southern California (USC), University College Dublin
  • Elsivier, Published: October 2011, Paperback, 672 Pages
  • The book provides comprehensive coverage of distributed and cloud computing, including:
    • Unordered List ItemFacilitating management, debugging, migration, and disaster recovery through virtualization
    • Clustered systems for research or ecommerce applications
    • Designing systems as web services
    • Social networking systems using peer-to-peer computing
    • Principles of cloud computing using examples from open-source and commercial applications

• An Energy-efficient Task Scheduler for Multi-core Platforms with per-core DVFS Based on Task Characteristics PDF
  • Ching-Chi Lin, You-Cheng Syu, Chao-Jui Chang, Jan-Jan Wu, Pangfeng Liu, Po-Wen Cheng and Wei-Te Hsu
  • September 2014 International Conference on Parallel Processing (ICPP-2014)

• Cost Minimization for Scheduling Parallel, Single-threaded, Heterogeneous, Speed-scalable Processors PDF
  • Rashid Khogali & Olivia Das
  • Department of Electrical & Computer Engineering Ryerson University, Toronto, Canada

ACE: Exploiting Correlation for Energy-Efficient and Continuous Context Sensing

• Suman Nath
• Microsoft Research
• MobiSys’12, June 25–29, 2012, Low Wood Bay, Lake District, UK
• ACE (Acquisitional Context Engine), a middleware that supports continuous context-aware applications while mitigating sensing costs for inferring contexts. ACE provides user’s current context to applications running on it. In addition, it dynamically learns relationships among various context attributes (e.g., whenever the user is Driving, he is not AtHome). ACE exploits these automatically learned relationships for two powerful optimizations.

CITA: Code In The Air Simplifying Sensing and Coordination Tasks on Smartphones

• Lenin Ravindranath, Arvind Thiagarajan, Hari Balakrishnan, and Samuel Madden
• MIT Computer Science and Artificial Intelligence Laboratory, Cambridge, MA, USA
• HotMobile’12 February 28–29, 2012, San Diego, CA, USA
• This paper introduced the CITA architecture for simplifying the tasking applications for users and developers. Developers, who can create tasks by writing only server side code, even for tasks that involve multiple end users and their devices, a variety of sensors, and the server. In our current implementation, these tasks are written in JavaScript. End users, who are able to specify their own tasks by “mixing and matching” available activities and tasks via a web UI (or a smartphone UI)

Snooze: Energy Management in 802.11n WLANs

• Ki-Young Jang, Shuai Hao, Anmol Sheth, Ramesh Govindan
• University of Southern California, Technicolor Research
• ACM CoNEXT’11, December 6-9, 2011
• The design and implementation of Snooze, an energy management technique for 802.11n which uses two novel and inter-dependent mechanisms: client micro-sleeps and antenna configuration management. In Snooze, the AP monitors traffic on the WLAN and directs client sleep times and durations as well as antenna configurations, without significantly affecting throughput or delay. Snooze achieves 30~85% energy-savings over CAM across workloads ranging from VoIP and video streaming to file downloads and chats.

COCA: Computation Offload to Clouds using AOP

• Hsing-Yu Chen, Yue-Hsun Lin and Chen-Mou Cheng
• National Taiwan University and Carnegie Mellon University
• 2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing
• COCA is a programming framework that allows smartphones application developers to offload part of the computation to servers in the cloud easily. COCA works at the source level. By harnessing the power of AOP, COCA inserts appropriate offloading code into the source code of the target application based on the result of static and dynamic profiling.

Collaborative Sensing over Smart SensorsPDF Just 2 pages? It is Workshop paper and contains 4 pages?

• Vassileios Tsetsos, Nikolaos Silvestros, and Stathes Hadjiefthymiades
• Pervasive Computing Research Group, Dept of Informatics and Telecommunications, University of Athens Panepistimiopolis, Ilissia,Greece.
• 2nd Student Workshop on Wireless Sensor Netwotks, Athens 2009
• IPAC (Integrated Platform for Autonomic Computing) adopts a novel and pragmatic approach to context-aware computing.

Enhanced Collaborative Sensing Scheme for User Activity RecognitionPDF Just Poster! Find full paper It is just Demo poster paper of two pages. No Full paper available.

• Yuki Nishida, Yoshihiro Kawahara and Tohru Asami
• The University of Tokyo
• SenSys’11, November 1–4, 2011, Seattle, WA, USA.
• Share raw sensor data among users, yield new information
 add information of noisy environment.

CoSense – A Collaborative Sensing Platform for Mobile DevicesPDF Fine, but just poster. Check full paper It need payment

• Samuli Hemminki,Kai Zhao , Aaron Yi Ding, Martti Rannanjärv
• Department of Computer Science, University of Helsinki, Finland
• SenSys '13 Proceedings of the 11th ACM Conference on Embedded Networked Sensor Systems
• Opportunistically distributes sensing tasks between familiar devices in close proximity.

Energy-efficient collaborative context sensing on mobile phonesPDF This is proposal, not paper! But check this paper OptiMuS

• Julien Eberle
• LSIR, I&C, EPFL
• EDIC RESEARCH PROPOSAL
• Explores optimized enablers that were developed to support context sensing on smartphones like continuous location tracking or activity recognition.

Sensing as a Service: A Cloud Computing System for Mobile Phone SensingPDF

• Xiang Sheng, Xuejie Xiao, Jian Tang and Guoliang Xue
• Syracuse University and Arizona State University
• SENSORS'2012: IEEE Sensors Conference; October 29-31, 2012, Taipei
• Identify unique challenges of designing and implementing an S2aaS cloud, review existing systems and methods, present viable solutions.

Crowdsourcing to Smartphones: Incentive Mechanism Design for Mobile Phone SensingPDF

• Dejun Yang, Guoliang Xue, Xi Fang
• Arizona State University
• Mobicom '12 Proceedings of the 18th annual international conference on Mobile computing and networking
• Designed incentive mechanisms for mobile phone sensing.

Mobile Sensing for Social CollaborationsPDF

• Chi Harold Liu, Pan Hui
• IBM Research - China and Deutsche Telekom Laboratories
• ACM CSCW Workshop of Mobile Collaboration in the Developing World, Hangzhou China, March 2011
• The main objective of this paper is to receive early feedback from researchers and practitioners in the areas of mobile computing and social collaboration, and to brainstorm and synergy new collaboration opportunities.

Energy-Efficient Sensor Node Control Based on Sensed Data and Energy Monitoring PDF

• Ho-Guen Song, Dae-Cheol Jeon, Hee-Dong Park
• Korea Nazarene University, Cheonan
• Springer-Verlag Berlin Heidelberg 2011
• Proposes an energy-efficient sensor node control mechanism to prolong sensor networks’ lifespan by minimizing and equalizing energy consumption of sensor nodes.

Energy-efficient Tasking in Participatory Sensing SystemsPDF

• Kevin Wiesner, Sebastian Feld
• Institute for Computer Science Ludwig-Maximilians-Universität München
• Location-based applications and services (LBAS) 2013
• Proposed to energy-efficient task distribution and monitoring concept for participatory sensing system and evaluate it by Means of simulation.

Wireless Sensor Network: A Promising Approach for Distributed Sensing Tasks PDF

• Prof. Madhav Bokare, Mrs. Anagha Ralegaonkar
• SSBES`s Institute Of Technology and Management ,Nanded, India
• Journal of Engineering Technology and Management Science Vol. I No.1 December-January 2012
• Most sensor networks actively monitor their surroundings, and it is often easy to deduce information other than the data monitored.

Sensing task assignment via sensor selection for maximum target coverage in WSNsPDF

• Marjan Naderan,Mehdi Dehghan , Hossein Pedram
• Amirkabir University of Technology,Tehran,Iran
• Assigning the sensing task to cover maximum number of targets while minimizing the energy consumption of the sensing operation.

Efficient Opportunistic Sensing using Mobile Collaborative Platform MOSDENPDF Good reference for platform design

• Prem Prakash Jayaraman, Charith Perera, Dimitrios Georgakopoulos and Arkady Zaslavsky
• CSIRO Computational Informatics Canberra, Australia
• 9th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing October 20–23, 2013 Austin, Texas, United States.
• To develop a platform that is autonomous, scalable, interoperable and supports efficient sensor data collection, processing, storage and sharing.

Opportunistic Collaboration in Participatory Sensing EnvironmentsPDF Interesting paper

• Niwat Thepvilojanapong, Shinichi Konomi , Yoshito Tobe, Yoshikatsu Ohta
• Tokyo Denki University
• MobiArch’10, September 24, 2010, Chicago, Illinois, USA.
• To achieve energy efficiency and reduce data redundancy, we propose Aquiba protocol that exploits opportunistic collaboration of pedestrians.

Dynamic Mobile Cloud Computing: Ad Hoc and Opportunistic Job Sharing pdf

• Niroshinie Fernando, Seng W. Loke and Wenny Rahayu
• La Trobe University, Australia
• 2011 Fourth IEEE International Conference on Utility and Cloud Computing, 281-286
• Explore the feasibility of a mobile cloud computing framework to use local resources to solve these problems. The framework aims to determine a priori the usefulness of sharing workload at runtime. The results of experiments conducted in Bluetooth transmission and an initial prototype are also presented.

To offload or not to offload: an efficient code partition algorithm for mobile cloud computing pdf

• Yuan Zhang, Hao Liu, Lei Jiao, Xiaoming Fu
• University of Göttingen, Germany, Tsinghua University, China
• 2012 IEEE International Conference on Cloud Networking, 80-86
• Proposed an efficient code partition algorithm for mobile code offloading. Our algorithm is based on the observation that when a method is offloaded, the subsequent invocations will be offloaded with a high chance. Unlike the current approach which makes an individual decision for each component, our algorithm finds the offloading and integrating points on a sequence of calls by depth-first search and a linear time searching scheme.

• Computation with parallelism and distribution and faults
• Input –> Map –> Interm. Output –> Reduce –> Final Output
• Input: (K, V)
• sample: counting word accurances
  • K: key: doc name, line no
  • V: value: contents of line
  • TODO by Nha: figure out how each reducer selectively process words that belong to specific range that it is responsible for.
    • Answer
  • Parallelism
    • mappers can run simultaneously
    • reducers can run simultaneously
    • but reducer can run after all mappers are done
  • Fault tolerance
    • master checks aliveness of each mapper/reducer, and re-executes the incomplete task

• Egalitarian distribution: reduce sampling rates of each device to get collective sampling rate of R_i
• Selective distribution: choose k nodes out of p_i+1 so that collective sampling rate of them meets R_i, unchosen nodes don't sample/report.
  • Centralized:
    • K-means algorithm to get k_i clusters, and choose one for each cluster
  • Distributed:
    • use probability of being representative
• Peformance metric
  • Sensing resolution
    • actual sensing rate / goal sensing rate, ignoring excessive sampling
• Energy saving percentage
  • (E-E')/E*100
• TODO for Nha
  • present their evaluation method & results

• Crowdsensing architecture
• TODO for Dileep
  • Study related work systems
  • Make comparisons between them, including MOSDEN
  • MCS Survey

• Goal: Monitor underwater using sound / pressure
• TODO:
  • How do we know the distance between sensors? Only care about vertical distance? Why?
  • How each node knows the coordinations of other nodes to detect dominating triangles?

• Understand more detail of the paper's algorithm, model, technique

• TODO:
  • how to enforce manhattan model's constraints.

• Check the following papers in the KAIST's wiki
  • Mobile cloud computing on 10/23 (postpone)
  • CUCKOO on 10/30 by Nha
  • Refactoring paper on 10/30 by Nha
  • Dynamic Mobile … on 11/06 by Dileep
  • COCA paper on 11/13 (postpone)
  • To offload or not on 11/20 by Dileep

• Java
• Use Activity/Service framework of Android
• QR-code/Resource Description file-based master-slave binding
• Use ibis frame work for RPC
• Question
  • How to decide the remote offloading ? –> next paper

• Offload by (groups of) classes