Referee 1 ******************************************************************

I think this is a well written paper, and recommend publish as is.

1) They did a good job motivating their work and giving a survey of related work.
2) They did a good job explaining the capabilities of their simulation library, with
    some detailed examples.
3) They also indicated where future work would be necessary.  This part was a
    bit thin. QoS (failures) and advance reservations were mentioned, but contract guarantees
    (or absence and violation thereof) were not mentioned. Nor was any attention paid to
    the emerging replicated global file repositories (maybe still too new). The most difficult
    problem may also be the scheduling priorities of a particular resource (which may have
    goals to service some users better than others).
4) Good list of references.
5) Fonts in Figure 1 are poor quality in the pdf file viewed onscreen 
   (probably should fix for pub)

Referee 2  ******************************************************************

Summary:

The paper attempts to describe GridSim which claims to be a modeling toolkit to simulate and evaluate scheduling strategies within Grids.


Technical contribution:

The technical contribution of this paper is low in the current stage as it only provides a simulation package that is at this stage not suited for Grids. The authors have rightly discovered the problems of a Grid, but fail to convince me that the simulation package they have can address these issues. It is not sufficient to write a simulation package, but one must also identify acceptable scenarios in a "World Wide Grid". The example given in the paper was a tiny local cluster problem. It is not convincing.

Thus the title of the Paper is totally misleading. Based on the title I was real excited about reading the paper, but was utterly disappointed at the end.

Before a publication in a journal can occur many theoretical and practical improvements to the work are necessary (networking, whereto obtain data to run a meaningful simulation, use of global instead of local, use of Grids instead of Cluster, ...)


Other comments:

I do not agree with the authors statement that the goals of resource brokers in Grids and supercomputing/cluster computing are different. Maybe this distinction is introduced to motivate the authors work. This needs to be rethought and explained more logically. The statement is given without proper explanation, and I doubt the authors will be able to proper explain it.

F: Presentation Changes


Format changes:

At times the paper is to wordy and uses a non academic language. The authors made the mistake to follow loosely the paper guideline given in the GCE working group, which made the paper somewhat more difficult to read A requirements analysis section is urgently needed. A better requirements analysis section is needed.

Changes

* Abstract: Shorten abstract, Several sentences seem redundant, first paragraph belongs in introduction. Do not explain a virtual organization in abstract.
* Abstract: In Grid environment"s"

* Introduction, The definition of a Grid is somewhat different from Fosters definition and the architecture presented in the "anatomy" paper. This should be pointed out.
* The four layers are not intuitive in the figure used to explain the architecture. Formulate the paragraph differently, so the reader does not try to look for the four layers.
* World Wide Grid, must be capitalized
* test test-bed seems redundant
* last sentence on page 2 seems redundant
* why point out that students investigate, aren't students also researchers?
* The author may need to investigate the use times while document writing. Why not write "We provide a " instead of we have proposed and developed?
* The term Gridlets needs to be defined better
* Section 3, do not just list the features. Explain what you can do with the features. Introduce a classification of features. Are these features not available in the related work projects? I like to see a comparison between the packages mentioned in related work in regards to the features list. Which features are provided by the other tools that your GridSim does not provide?
* Remove the colloquial word "basically" from your vocabulary and replace it with a meaningful term.
* The term define is not used properly. You "define" something, but it is impossible to say "basically define."
* Section 3.3,  needs significant rewriting as no one knows about the simjava classes. This information will be largely useless to the reader. It is unclear in which way the classes are extending the JavaSim package. Even after reading section 3.3 several times I am unable to identify the author changes/contributions. It must be clearly spelled out what was provided by JavaSim and what was extended by GridSim.
* Make GridSimTags a Figure.
* Section 4, starts with redundant sentence, you have said this be now 5 times.
* Check the whole paper to use active voice instead of passive voice.
* The usage of the word "believe" is not recommended in scientific publications.
* The definition of Gridlets occurs 10 pages after it was first used.
* The reference to Figure 4 and its explanation in a separate section is confusing.
* Do not use colloquial terms like "for the sake of"
* The explanation why you use Java is unimportant other than the use of oject oriented technology.