The costs and benefits of turning data into information using big data systems

Abstract

Bu tez kurumsal rmalardaki buyuk miktarlardaki verinin, islenmeside ve saklanmasndaki zorluklar ve cozumlerini acklmaktadr. Yeni ortaya ckan buyuk verikavramnda isletimsel, mimari ve kullanmsal olarak bircok problem bulunmaktadr.Temel veri isleme yontemleri yeni olmamasna karsn, olusan veri miktar ve hz,merkezi mimarilerin snrlarn zorlamaktadr. Hadoop gibi, dagtk isleme modellerinikullanan, dagtk mimariler gunumuzde buyuk verinin islenmesi ve saklanmasnda anahtar rol oynamaktadrlar. Bu tez buyuk veriyi islemede, yapsal, mimarive nansal olarak sorun teskil eden durumlarn gercek hayat deneyimleri uzerinden,telekominikasyon, nans ve petrol ranerileri gibi sektor bazl cozumlerini acklamayamaclamaktadr.

This thesis explains problems of and solutions for storing and processing big volumesand streaming types of data in a cost-eective way for enterprise companies. Thereare several problems like operational, infrastructural and usability problems aboutthese new concepts of Big Data. The basic data processing concepts are not new,but the data generation volumes and velocities are pushing the limits of centralizedarchitectures. Distributed systems using distributed programming models such asthe Hadoop framework are used today to handle big data problems. This thesis willtry to combine the structural, architectural and nancial issues to address big datastorage and processing problems and will give practical examples based on real-lifeexperiences from several big data applications in dierent sectors including mobiletelecommunications, nance and oil-gas elds.