1. Course Abstract

Purpose : Though it is only about fifteen years since the word "Grid" was introduced in computer science, a lot of research efforts have been pursued across the world. Grid computing has evolved in various directions and matured part of this progress will be captured in this course. This course will introduce the students fundamental concepts in Grid and move on to Cloud computing for various applications. We will dwell on the fundamental issues of resource discovery, allocation, etc before moving on the pricing issues in Cloud services. This course will introduce some fundamentals in cloud architectures, modules that make the resources provision work seamless. Many of the current Cloud services, SaaS, IaaS, PaaS leading to XaaS will be introduced and discussed in class. The key behind-the-door concept in Cloud, Virtualization, will then be discussed followed by tools for hardware virtualization such as VMWare ESXi, Xen, KVM. Customer demands for self-service, per-usage metering and billing, Elasticity will be touched upon, which will lead the class to pricing issues in Cloud. This course will introduce the students some fundamental concepts from finance to design a mechanism for pricing resources. Some applications from Science, Engineering and Finance will be introduced and the current trend in the use of Cloud resources for High Performance Computing will be introduced. The following is an outline of the topics to be covered and evaluation to be adopted.

2. Topics covered

in this course will include:
1. Introduction to Grid Computing:
1. History, Infrastructure development, Applications and Software interface
2. Diverse Policies, Regulations, Security and Risk Management in Grids
3. Commercial Grid Services (Amazon S3, EC2, GoGrid ...)
4. Current Grids: Canadian, US, Europe and rest of the world. Some emphasis will be placed on the WestGrid in which Univ. of Manitoba is a constituent; Students will be required to get accounts on WestGrid infrastructure for their use during the course and for doing their projects later on.
5. Science, Engineering and Finance applications on Grid (general discussions).
2. Virtualization and Cloud
3. Grid/Cloud Architecture and Execution Models
1. Globus, Gridbus, Condor, Nimrod-G, EGEE (Enabling Grid for E-science in Europe)
2. Layers and Types of Cloud
3. MapReduce (Hadoop)
4. Grid Portals
1. Grid service deploying techniques.
2. Students will be introduced to write simple deployment descriptors. This will also help in preparing them for (F (a) and (b)) below.
5. Ownership and Decentralization
6. Grid Simulators (GridSim will be discussed in class)
1. Students will use simulation environment (e.g., SimGrid, GangSim, OptorSim) to do their projects and may migrate to a real Grid based on the type of their project.
7. Resource Discovery (Brokers)
8. Resource Management
1. Reservation based allocation
2. Auction based allocation
3. Economy based schedulers
4. Nimrod/G Resource broker and scheduling
5. Game theoretic scheduling
6. Quality of Service (QoS) guarantee
7. Job scheduling
9. Pricing Cloud Resources: Economic and Finance Models
10. Running Applications on Grid/Cloud: Case Studies from Science, Engineering and Finance. Couple of real problems will be discussed.
A detailed weekly schedule is available at:

Course Schedule

3. Recommended Readings:

There is no single book that covers all the topics in its entirety. The books below are recommended for reference purposes. I am preparing my own notes from various sources and making them available on the course web site along with many relevant and current papers (journal and conferences) and book chapters as hand-outs.
1. The Grid -- Blueprint for a new computing Infrastructure, Ian Foster and Carl Kesselman, Morgan Kaufmann, 1999.
2. Grid Computing: Techniques and Application, Barry Wilkinson, CRC Press, 2010
3. Market-Oriented Grid and Utility Computing, Raj Buyya and Kris Bubendorfer, Wiley, 2009.
4. Cloud Computing: A Practical Approach, A.T. Velte, T.J. Velte and R. Elsenpeter, McGrawHill, 1st edition 2010.
5. Cloud Computing: Principles and Paradigms , Rajkumar Buyya, Jame s Broberg and Andrezej Goscinski (Ed.) Wiley, 2011.
6. Lecture slides, Handout documents, on-line materials, Ruppa K. Thulasiram
7. Papers from very recent conferences.
8. Papers from recent journal issues.

4. Evaluation

1. Evaluation Scheme:

   The breakdown of the evaluation scheme is as follows:

   Assignments	20%
   MidTerm	30%
   Final Project	50%

2. Assignments:

   Will include mini projects such as
   1. design of web based portal
   2. installing Grid simulation tool kit(s) (at least one)
   3. job scheduling in the simulation environment
   4. application development in the real or simulation environment
   5. ONE of the assignments (A2) will be on paper presentation. Each student will read TWO papers (one each from Grid and Cloud) and present one of the papers. These two papers will be from two different areas of Grid and Cloud (architectures, applications, benchmarks, execution, simulators, scheduling, utility, virtualization etc.). For example, if a student selects his/her first paper from Grid Architecture, his/her second paper has to be from Cloud Computing literature other than Cloud Architectures area.
      There will be TWO readers for each paper presented in class, who will be asking a minimum of THREE questions primarily and check if the core of the paper is well presented by the speaker. Each student will write summary (one-page each) on the two papers they questioned/critiqued and will also fill-in a questionnaire on the presentation by the speakers of these two papers. While filling in the questionnaire with mostly Yes/No answers, the student will also mention his/her question and speaker's response and if the answer was appropriate.
      In addition, each student will write a thorough review on the second paper they selected to read.
      In essence, each student will read 4 papers in total [2 papers selected by the individual (one for presenting and one for summary review) and 2 papers assigned by the instructor from the selection by rest of the class (for questioning, writing one-page summary on each of these papers and filling-in questionnaire on each of the two presentations)]

3. Mid Term Exam:

   It will be in-class lasting 75 min testing student's knowledge on concepts, current systems, and applications. Given the current commercial Cloud scenario, one of the questions will explore their thoughts on possible future directions.

4. Course Project:

   You will be given an opportunity to come up with ideas for course project, which will be discussed and sorted out at appropriate time during the course. Or, you can also select a project from a list of topic areas given below. This list will be suggested to a student only if he/she cannot come up an idea for the project.

   Project Topic Areas:

   1. Resource allocation
   2. Resource scheduling algorithms and techniques
   3. Job scheduling
   4. Business/Enterprise Grid models
   5. Clouds
   6. QoS mechanisms
   7. Pricing mechanisms: economic models
   8. Pricing mechanisms: finance models
   9. Load Balancing in Grid
   
   The project will consist of the following tasks:
   1. Project Selection: Before Feb.10, 2012 all the students will be required to submit a proposal on their topic of interest after individual discussion on general features.
   2. Proposal presentation: They will be required to give a brief presentation on their proposal in class or in-person to the instructor (before Feb.17, 2012 ) soon after selecting a topic for the project. They will also be required to provide a write-up of their proposed work describing
   1. the problem
   2. challenges
   3. approaches
   4. expected results
   I will discuss with the student on the goal, outcome and especially on the feasibility of methodology proposed for the project as described in the proposal. The coding has to be done in the most appropriate environment.
   3. Project work: They will be working on their projects individually, reporting to me or consulting with me on a weekly basis.
      If a project involves a group of a maximum of two students a CLEAR distinction has to be established on the work and load has to be distributed equally between the two in the group.
   4. Deliverables: They will present the partial completion report to the class (on a special class) or in-person to the instructor and hand in a report together with the code they developed for the project by the last day of the classes (that is, by April 5, 2012).
      1. Project Report: A following template will be provided to the students to maintain uniformity of the report and to help me judge their work in the given limited time of the term.
         1. Title, Abstract page and Project Number (will be assigned)
         2. Introduction
         3. Problem statement
         4. Literature review
         5. Proposed solution (approach) Results and discussions
         6. Possible extensions
         7. Conclusions
         8. References
         9. Appendix: Code developed
      2. Project Presentation with implementation details in the same order as the report.
      3. Project work Evaluation: Rate the significance of the report on a scale of 5 (incremental contribution), 4 (no new contribution, however thorough implementation of existing concepts/ideas), 3 (insufficient effort), 2 (no effort), 1 (no understanding of the concepts).
      4. Above ratings will be reached by asking the following questions among others
         1. how convincing the methodology is. You may consider some of the following questions (use what is relevant):
         2. objectives achieved?
         3. realistic assumptions?
         4. experiments well designed?
         5. can different experiments be designed?
         6. alternatives methodology suggested?
         7. limitation of the approach understood?
         8. Results and discussions
         9. Possible extensions
         10. Conclusions
         11. References
         12. Appendix: Code developed

Instructor

Lecture Timings: