Spatial Cloud Computing: How can the geospatial sciences use and

advertisement
Chaowei Yang , Michael Goodchild , Qunying Huang ,
Doug Nebert , Robert Raskin , Yan Xu , Myra Bambacus &
Daniel Fay (2011) Spatial cloud computing: how can the
geospatial sciences use and help shape cloud computing?,
International Journal of Digital Earth, 4:4, 305-329,
Presenters: Gayathri Gandhamuneni, James Wang
Team URL: http://www-users.cs.umn.edu/~yumeng/
Topics
 Motivation
 Problem Statement & Illustration
 Challenges
 Major Contribution
 Validation Methodology
 Proposed Approach – SCC Scenarios
 Key Concepts
 Cloud Computing, Spatial Cloud Computing
 Assumptions
 Preserve and Revise
Motivation
 Constant changes
 Better recorded through space – time dimensional data


Exabytes of data accumulated
Increasing at rate of PB
 Analysis of information changing


Understand, protect & improve living environment
Ex: Predict events like earthquakes, tsunamis…
 Need of computing infrastructure that can
 Reduce IT work
 Real time applications support
 Deal with access spikes, Support massive users
 System of System Solutions
Problem Statement
 Input: Geospatial Sciences (GS) Information
 Output: Computing Infrastructure suitable for GS
 Objective: Research on challenges in geospatial sciences
and use of Spatial Cloud Computing for solutions.
 Constraints: SpatioTemporal Principles & Geospatial env.
Challenges
 Information Technology challenges for Geospatial sciences
 Data Intensity

Support of massive data storage, processing & system expansion
 Computing Intensity


Algorithms and models based on Earth phenomena are complex
Complexity grasp of spatiotemporal principles
 Concurrent Access Intensity

Lot of end users trying to access concurrently
 Spatiotemporal intensity


Geospatial datasets  space – time dimensions
Spatiotemporal – Static/Dynamic
Major Contributions
 Categorization - Challenges of Geospatial Sciences in
21st century
 Relation of Cloud Computing & Geospatial Sciences
 Cloud Computing usage and how spatiotemporal
principles enhance it
 Examples to show how spatial cloud computing can
solve 4 intensity problems


Most Significant
Looks ahead to see possible solutions for intensity problems
Cloud Computing
 Advanced Distributed Computing
 Provides ‘computing as a service’
 ‘Pay-as-you-go’ model
 Model:
 Convenient, on-demand network access
 Access to shared pool of computing resources

Ex: networks, servers, storage, applications and services
 Resources can be provisioned and released fast


Minimal management effort
Service provider interaction
Characteristics of Cloud Computing
 Cloud Computing difference to other distributed approaches
 On-Demand Self Service

As needed automatically
 Broad Network Access

Different types of network terminals
 Resource Pooling

Consolidation of diff. types of Computing resources
 Rapid Elasticity

Rapidly & elastically provisioning, allocating & releasing resources
 Measured Service

Supports pay-as-you-go approach
Advantages of Cloud Computing
 Rapid Deployment
 Dependability/Redundancy
 Flexibility/Scalability
What are the advantages of Cloud Computing?
 Levelled Playing
Field
 Security
 Identity Management & Access Control
Services for Cloud Computing
 Cloud Computing is provided through 4 services
 Infrastructure as a Service (IaaS)
 Platform as a Service (PaaS)
 Software as a Service (SaaS)
 Data as a Service (DaaS)
Geospatial Sciences
Uses of Cloud Services
 Earth Observation (EO) Data Access:
 Fast, secure access & utilization of EO data
 Storage & Processing needs - DaaS
 Parameter Extraction:
 Complex geospatial processes – Reformatting & Reprojecting
 PaaS can be used
 Knowledge & Decision Support:
 Used by domain experts, managers or public
 SaaS provides good support
 Social Impact & Feedback:
 SaaS such as Facebook & email can be best utilized
Spatial Cloud Computing (SC2)
 Cloud Computing Paradigm
 Driven by geospatial sciences
 Optimized by Spatiotemporal principles
 Geospatial Science Problems
 Intensive Spatiotemporal constraints & Principles
 Best if spatiotemporal rules for geospatial domains used
GeoSpatial Principles
 Physical phenomena are
 Continuous
 Heterogeneous in space, time, and space-time scales;
 Semi-independent across localized geographic domains and
can be divided and conquered
 Geospatial science and application problems include the
spatiotemporal locations of
 Data Storage
 Computing/processing resources
 Physical phenomena
 Users
 Spatiotemporal phenomena that are closer are more
related (Tobler’ first law of geography)
Spatial Cloud Computing Framework
Validation Methodology
 Four scenarios given for 4 intensity problems in order
to validate their work
 Case study to show that SCC might solve the four
problems of geospatial sciences
SCC: Data Intensity Scenario
SCC: Computing Intensity Scenario
SCC: Concurrent Access Intensity Scenario
SCC: Spatiotemporal Intensity Scenario
 Real-time traffic network - Metropolitan area like DC,
 Static Routing – 90k nodes, 200k links, 90k*90k origin
& destination requests

Several Optimized routes for one OD request pair – 1 GB
 Dynamic Real – Time Routing




Routing condition – Changes for each min. and each link &
node
Daily - Volume increases by about (2460) 1TB
Weekly– (24607) 10TB
Yearly - (2460365)- 1PB
Assumptions
 Methods and principles of geospatial sciences that can
drive and shape computing technology would remain
unchanged
 Unreliable assumption
 Both the development in technology & geospatial sciences
itself might cause changes to occur
 Validation done with examples of particular scenario
 Can cloud computing be used always
 Overhead cost of cloud computing might be > Cost without
cloud computing
Application Areas
 Spatiotemporal principle mining & extracting
 Important digital earth & complex geospatial science
and applications
 Supporting the SCC characteristics
 Security
 Citizen and Social Science
Present & Future
 Present:
Present & Future
 Present:
 Google Maps: Encouraged Web developers
 Other Companies: GISCloud.com, SpatialStream.com


Web based solutions for GIS functions
Spatial Analysis & Data management
 ESRI’s ArcGIS Online – ArcGIS.com
 Future:
 Security – Personal & Sensitive data
 Boundaries



Mostly on internet
Wary about location of data and services
Source: http://www.linkedin.com/groups?gid=1839124
Exercises/Questions to Check
 What are the problems faced by geospatial data?
 What are geospatial principles?
 What does system of systems solution include?
 What is Cloud Computing?
 Different services of Cloud Computing?
 How is Cloud Computing different from others?
 What is Spatial Cloud Computing?
 What scenarios Spatial Cloud Computing can be used in
context of geospatial sciences?
Preserve & Revise
 Revise
 Whole paper - Recent advancements in cloud
computing
 More practical examples of SC2 scenarios
 Security issues faced and any possible solutions
 Preserve
 Different types of intensities
 Cloud Computing & SC2 key concepts

Relationship between both
References
 [1] Chaowei Yang , Michael Goodchild , Qunying Huang , Doug Nebert , Robert Raskin , Yan Xu
, Myra Bambacus & Daniel Fay (2011) Spatial cloud computing: how can the geospatial
sciences use and help shape cloud computing?, International Journal of Digital Earth, 4:4, 305329, doi: 10.1080/17538947.2011.587547
 [2] Buyya, R., Pandey, S., and Vecchiola, S., 2009. Cloudbus toolkit for market-oriented cloud
computing. Cloud Computing, Lecture Notes in Computer Science, 5931 (2009), 24_44. doi:
10.1007/978-3-642-10665-1_4.
 [3] Olson, A.J., 2010. Data as a service: Are we in the clouds? Journal of Map & Geography
Libraries, 6 (1), 76_78.
 [4] Mell, P. and Grance, T., 2009. The NIST definition of cloud computing Ver. 15. [online].
NIST.gov. Available from: http://csrc.nist.gov/groups/SNS/cloud-computing/
 [5] Yang, C., et al., 2011a. WebGIS performance issues and solutions. In: S. Li, S. Dragicevic,
and B. Veenendaal, eds. Advances in web-based GIS, mapping services and applications.
London: Taylor & Francis Group, ISBN 978-0-415-80483-7.
 [6] Yang C., et al., 2011b. Using spatial principles to optimize distributed computing for
enabling physical science discoveries. Proceedings of National Academy of Sciences, 106 (14),
5498_5503. doi: 10.1073/pnas.0909315108.
THANK YOU
Download