Analysis of Remote Sensing Quantitative Inversion in Cloud

Analysis of Remote Sensing Quantitative Inversion in Cloud Computing Jing Dong, Yong Xue, Ziqiang Chen, Hui Xu, Yingjie Li Institute of Remote Sensing Applications Chinese Academy of Sciences July 29, 2011

• Aerosol Optical Depth(AOD) Retrieval • High Performance Computing Institute of Remote Sensing Application, CAS

Outline § An Introduction of Cloud Computing § Remote Sensing Quantitative Inversion & Cloud Computing § A Possible Cloud In Remote Sensing Quantitative Inversion § Discussion and Future Work Institute of Remote Sensing Application, CAS 1

Introduction § What is Cloud Computing § What can Cloud Computing do § How does Cloud Computing work Institute of Remote Sensing Application, CAS 2

Introduction A large-scale distributed computing paradigm that is driven by economies of scale, in which a pool of abstracted, virtualized, dynamically-scalable, managed computing power, storage, platforms, and services are delivered on demand to external customers over the Internet. ——Foster et al. (2008) • Saa. S Software as a Service • Iaa. S Infrastructure as a Service • Paa. S Platform as a Service • Daa. S Data as a Service Institute of Remote Sensing Application, CAS 3

Introduction-cloud products Amazon Elastic Compute Cloud (Amazon EC 2) of Amazon、Windows Azure of Microsoft、 Google App Engine of Google、Blue Cloud of IBM Institute of Remote Sensing Application, CAS 4

Remote Sensing Quantitative Inversion & Cloud Computing § Charge eg. Amazon EC 2 § Technical Support eg. Google App Engine § Security computing infrastructure maintained and controlled by third parties Institute of Remote Sensing Application, CAS 5

An example-Charge § Application—remote sensing quantitative inversion § Data Resource ：MODIS for one year § Range : over China § total volume: 2. 97 TB (2008) § per day: 8. 4 GB Institute of Remote Sensing Application, CAS 6

Charge Standard of Amazon in US dollars-data transfer nearly 304. 13 US dollars Institute of Remote Sensing Application, CAS 7

An example-Charge § Data Resource ：MODIS for one year § Range : over China § total volume: 2. 97 TB § per day: 8. 4 GB § processing time of geometric correction(GC) Institute of Remote Sensing Application, CAS 8

Charge Standard of Amazon in US dollars The price of on-demand instances at least 832. 2 US dollars Institute of Remote Sensing Application, CAS 9

An example-Cost § Data transfer in— nearly 304. 13 US dollars for the whole year of 2008 § Instance rent for GC— at least 832. 2 US dollars for one year data The cost is so high Institute of Remote Sensing Application, CAS 10

An example -Technical Support § Google App Engine • Java • python § most of the quantitative remote sensing inversion algorithms • IDL • C++ Institute of Remote Sensing Application, CAS 11

Security § Data Security § Network connection § Emergency • breakdown in 2008 and 2011 • service closed Institute of Remote Sensing Application, CAS 12

A Possible Cloud Advantage: avoid high cost for data transfer Disadvantage: not suitable for multi-source data comprehensive applications Institute of Remote Sensing Application, CAS 13

Discussions and Future Work § Cloud computing would not necessarily suitable for all scientific applications, especially for geoscience § Standardized metadata management mechanism, more efficient data storage and sharing method, and more reliable security mechanism are needed Institute of Remote Sensing Application, CAS 14

Thank you Jing Dong Email: dongjing 0311@hotmail. com URL: www. tgp. ac. cn