Stephanos Androutsellis PhD research plan

   Publications Plan:

• Currently I only focus on thesis write-up. There are no specific goals for other publications.

   Non-thesis-related personal plans:

• Follow SENSE/PENED accounting stuff.
• Follow SOPRANO accounting stuff.
• Prepare for SOPRANO Bilbao meeting (June 10-12).

PhD thesis table of contents:

   *1.	Introduction
   **a. Overview of my research and motivation

Use Ptrim S1

• • b. Main results and original contributions
  • c. Organisation of thesis
• 2. Theory and background
  • a. Introduction – big picture
  • b. P2P networks and architectures

Based on ACM survey

• • c. Distributed transaction processing systems

Compared to centralized, characteristics, requirements (use e-commerce paper)

• • d. Centralized and distributed reputation management

Survey of current approaches

• • e. Market-based risk management

Financial derivatives, risk management, market efficiency. Use Ptrim S2.1. S2.4

• • f. Auxiliary theory and tools used in this research

Mention, with references to Sections, we’ll be talking about crawling utilities, importance sampling, Java/jxta development etc.

• 3. Problem definition and proposed solution
  • a. Definition of problem we want to attack: Managing P2P transaction risk

Includes general and also specific goals

• • b. Limits of research scope

Things that are not our problem ( References to Section 7).

• • c. Proposed solution: A Market based approach

At an abstract level. Use patent literature for ideas.

• • • i. Relation to other approaches

Complementary to reputation management, explain how the two could cooperate.

• • • ii. Abstract description of proposed solution.

Abstract design and architecture. Description of the market-based insurance layer, and the Ptrim system. Use Ptrim S2.3. Talk about need for efficient, competitive market. Use Ptrim S3.1.

• 4. Methodology
  • a. How we chose this particular approach

With particular focus to our evaluation approach. Discuss other possible approaches (e.g. theoretical) that were not selected.

• • b. Description of our overall approach

We’ll build a prototype and test it, and we’ll also run composite experiments and evaluate the results.

• • c. Hypotheses

What were our specific hypotheses (H1, H2, H3). We’ll refer to them in d and e below, and also again in Section 6.

• • d. Prototype system trials

What sort of trials we ran with the prototype. Use Ptrim S4.1. Which hypotheses do we address here.

• • • i. Experimental setup

How the experiment will work, software tool, subjects, questionnaires etc.

• • • ii. Methodology for evaluation

How will we evaluate the results.

• • e. Composite experimental trials

Use Ptrim S4.2. Which hypotheses do we address here.

• • • i. Experimental setup

How the experiment will work, software tool, subjects, questionnaires etc. Use Ptrim S

• • • ii. Transaction data

Where we found the data and what we did with it

• • • iii. Methodology for evaluation

How will we evaluate the results (market efficiency, risk reduction, questionnaires etc)

• 5. Design and implementation

Follows the structure of 3c, but not at an abstract level any more, we refer to design and implementation. Lots of UML diagrams

• • a. Design of Ptrim system
  • b. Ptrim implementation details

How we implemented this: Java, jxta, messaging, etc Use Ptrim S3.2.

• • c. Composite experiment setup

How we implemented this: Perl/CGI/MySQL etc.

• • d. Data crawling

How we did this

• 6. Results, evaluation and discussion
  • a. Evaluation of prototype trial outcomes

Use Ptrim S5.1. Refer to hypotheses

• • b. Evaluation of composite trial outcomes

Use Ptrim S5.2. Refer to hypotheses

• • c. Discussion of overall findings and implications for our approach
• 7. Future work

Mention these also in Section 3.b.

• • a. Security
  • b. Collusion detection
  • c. Insuring insurance transactions
  • d. Further experiments and simulation studies
• 8. Conclusions and synopsis

PhD thesis research plan:

• May 29th 2008: Draft version of PhD skeleton, official initiation of writeup
• June 10th 2008: Completion of additional composite experiment with uniformly sampled data on undergrads. This will be used for the thesis, or for future Ptrim paper versions.
• September 2008: Expected revision of Ptrim journal paper.
• July 10th 2008: First draft of thesis for overview with dds.
• August 25th 2008: Second draft of thesis for overview with dds.
• September 20th 2008: Thesis version ready for review with dds.
• October 15th 2008: Final thesis version ready for submission.
• November 2008: Thesis presentation and PhD completion.