0

I am writing my first research paper on distributed storage systems. We also have a prototype working (partially). The project was a complete implementation based project where we envisage to take this project a step further and do some future development.

So, my questions are:
1. How should I go about creating the "skeleton" or the outline of the paper? Any pointers.
2. I understand, benchmarks and some working stats of the systems is important, but what is the breakup? How much content should be dedicated to theoretical algorithms (which we implemented, did not create new ones) and the benchmark data.
3. What format (IEEE, ACM etc) should I select and why?

PS: The project does not really concentrates on creating something new and unique, but it's aim to create an app which can be used by students and researchers to help understand the field of study better. Performance is not an objective for us, since there are alot of open source implementation (production quality) of the same domain.

Think of Kaffe for JVM, that's what we are building for distributed storage systems.

zengr
  • 1,217
  • 11
  • 22

1 Answers1

2

Usually research paper means you want to publicate it on an accepted platform, usually these are conferences. If so you should select such an conference, which has soon a submission deadline. Usually this conference dictate the layout of the submissions (so this answers your question 3.). From the other 2 question I would conclude that you have almost no paper experience. If you want to submit such things you should have a mentor (e.g. for phd-students this is there professor) who should answer such questions. Esp. because certain communities have certain demands which are known by experienced researchers. An general outline of such a paper is Abstract, Introduction, Prior/Related Work, Idea, Implementation, (Experimental) Results, Conclusion & Future Work (This should answer 1.). For 2 I could just say: Search someone experienced from your field, as different research communities have different expectations.

In general a not-research-paper (as yours is, as you said, you did nothing new) is unlikely to get accepted. In my field of research (compilers and hardware) the acceptance rate of the conferences or journals is 20-30% (of the submissions). And again: here an advisor would be really helpful.

If you write your paper, just to put it on web on a internal company server, use my above structure as guideline and weight the single points as you want.

flolo
  • 146
  • 2