40 likes | 170 Vues
This paper discusses the limited discourse surrounding current computational infrastructures and programming paradigms in the context of dynamic data applications. It questions why there are few complaints about the computational infrastructure available, particularly concerning scalability amid the anticipated surge in data. Despite the awareness of approaches like MapReduce, their adoption appears limited. The study emphasizes the classification of data-intensive applications and the identification of fundamental characteristics that can inform the understanding of programming paradigms for dynamic data utilization.
E N D
Dynamic Data & Programming Paradigms mark birkin, joelsaltz, pauljefferys, geoffreyfox, jeremycohen, omerrana, dankatz, neilchue-hong, ally hume, shantenujha
Why is everyone so “happy”? [Except Us] • Try to analyse why there have been limited complaints about the CI and PP available? • E.g only Szalay talked about Graywulf. Why? Gene Sequencers… • When people talk about the data deluge -- do your current approaches scale to the expected increase in data? • There has been no discussion of scalable approaches so far -- is it because this is not an issue for many people. • Why have people not talked about MapReduce?
In a nutshell: What is needed? • Neil: Is there a classification of data-intensive applications thatwould help me to understand (simply) if mapreduce were/were-notuseful. Involves: • (i) Classification of application charecteristics • (ii) Identification of “fundamental vectors” • (iii) Understanding the Programming Paradigms based upon (i) and (ii)
Dynamic (DI) Applications • What are the different classes of Dynamic DI Applications? • Real-time (sensor) data assimilation • Dynamic resource utilization • Application-level changes • Feedback eg. Sensor selection • Similar questions for to slide (ii) now with dynamic data-intensive