w 22 | ocus cnrsI InternatIonal magazIne F BIG DATA InThE humAnITIEsAnD soCIAl sCIEnCEs 01 “Big Data has revolutionized “When the information is fathom? How will such the work of specialists in the not collected directly by the uncontrollable humanities and social researcher, it is difficult to exhaustiveness impact our sciences,” indicates Bertrand know how the data was relation to knowledge?” Jouve, mathematician and processed before its inclusion There are a number of other deputy scientific director at in the database. LIG2 concerns, such as data the CNRS’s Institute for researcher Sihem Amer-Yahia ownership, usage rights, the Humanities and social believes this to be Big Data’s right to be forgotten, or Sciences (INSHS).1 He sees Big Achilles’ heel. “Raw data ethics. Researchers in the Data as a great opportunity processing is often a black humanities and social for his peers. “Online box, completely opaque sciences must address these databases now provide a to the user,” she explains, issues, in collaboration with single entry point for “yet it is a known fact that other disciplines, in order to cture accessing knowledge that common manipulations can serve the public interest and PI was previously scattered delete a large part of the avoid the possible I la across many locations,” he information.” stranglehold by private P tock/ adds. “Internet-based The emergence of very large interests. Ys surveys, for example, have data volumes and the aller 01.Institutdesscienceshumaineset not only made the work of all-digital world raises other sociales. g sociologists easier, but also issues, albeit less technical. 02.laboratoireinformatiquede © 01the digital era simplifies access to information given their research more “Big Data is inevitably a cause grenoble(cnrs/universitéde previously scattered across various libraries. reach.” Despite his for epistemological concern,” polytechniquedegrenoble). Institut/-IIIand-II,renoble-I,g enthusiasm, Jouve is also notes Sandra Laugier, deputy allocating $200 million to the country’s aware of the many difficulties scientific director at INSHS. contactInFormatIon: Big Data Research and Development facing users today. “The “What does it mean to have > bertrand.jouve@cnrs-dir.frBertrand Jouve Initiative. The EU has made management problem is how to process access to more information sandralaugier of digital content a priority for the end of the raw data,” he explains. than a human mind can > sandra.laugier@cnrs-dir.fr its 7th Framework Programme (FP7). In France, a €25 million program is dedica- Mathematics and High Performance amounts of scientific data. “What is the ted to big data management technologies. Computing at the French Ministry of best way to store and preserve data? How Higher Education and Research. This can it be processed, analyzed, viewed, sCIEnTIFIC ChAllEnGEs AhEAD prompted CNRS to launch the Mastodons and interpreted? How should it be pro- “Big Data is a considerable scientific chal- program last summer (see box p. 21). The tected, in particular from abusive use, lenge that can only be met by a combina- idea is to support interdisciplinary proj- and how can it be permanently deleted? tion of basic science and engineering,” ects that will identify the problems in- All issues that need to be addressed, and explains Mark Asch, scientific officer for volved in the management of very large for which we have few answers,” says ComPArATIvE sCAlE oF ByTEs Basic One page A piece of A two-hour six.ltab million A stack of All the All the data Storage capacity of unit of of text music film books DVDs information recorded the NSA datacenter measurement as tall as generated in two.ltabzero.ltabone.ltabone.ltab (nine.ltabtwo.ltab,zero.ltabzero.ltabzero.ltab m², two.ltabzero.ltabone.ltabthree.ltab) three.ltabzero.ltab KB five.ltab MB one.ltab GB one.ltab TB a five.ltabfive.ltab-storey up to two.ltabzero.ltabzero.ltabthree.ltab one.ltab B building one.ltab,eight.ltab ZB one.ltab YB five.ltab EB one.ltab PB Byte Kilobyte Megabyte Gigabyte Terabyte Petabyte Exabyte Zettabyte Yottabyte KB MB GB TB PB EB ZB YB cnrs one.ltabzero.ltabzero.ltabzero.ltab Bytes one.ltabzero.ltabzero.ltabzero.ltab KB one.ltabzero.ltabzero.ltabzero.ltab MB one.ltabzero.ltabzero.ltabzero.ltab GB one.ltabzero.ltabzero.ltabzero.ltab TB one.ltabzero.ltabzero.ltabzero.ltab PB one.ltabzero.ltabzero.ltabzero.ltab EB one.ltabzero.ltabzero.ltabzero.ltab ZB / IF atag D ©
CIM28
To see the actual publication please follow the link above