Communications of the ACM

Globus Online manages fire-and-forget file transfers for big-data, high-performance scientific collaborations.

Anonymous

Hi!
I am thankful to all authors of this good article, though its title has nothing to do with the content. The article itself is only about the data-movement; this purpose is clearly announced at the beginning of the second page and the whole text gives a lot of information about the declared subject: We report here on a research data-management system focusing on GOs data-movement functions.
Software for data scientists means a set of programs which help scientists to look through and to analyse those huge amounts of data; at least, this is my understanding of the underlined term. The whole life cycle of data can be divided into three stages: obtaining data, transferring it to another location (for one of the reasons which are mentioned in the article), and its analysis. This article describes some results on an interesting work on the second stage and by the way mentions the software for analysis as a service.
For those who provide the software as a service there are certainly big economical pluses; for scientists, who use such service, there are no pluses at all. Analysis of data and especially analysis of huge amounts of data is so individual that it requires absolutely individualized programs. For this reason nearly every laboratory, institute, or university develops a lot of its own software. Individualized programs must be developed in a different way from currently used; the new programs are not more costly in design but they are different in use. These user-driven applications give scientists the freedom in organizing their work; the same level of freedom which we had years (or decades) ago while working with paper and pen but now all benefits of computers are added to it. I wrote about it in a small article in CiSE (July/August 2011, v.13, N4, pp.79-84) and with much more details in a book; article contains a link to this book.
Data transferring and data analysis are different problems but strongly connected by this data flow. Usually the same people deal with both problems so I though that it might be interesting to them to know about the new results in the related area; these results can help them in their work.
Thank you for interesting results in data transferring area.
Dr. Sergey Andreyev,
Department of Mathematical Modelling, Heat and Mass Transfer Institute