Terradue is an Italian SME focused on providing cloud services for earth science research. They have developed an open platform to help scientists access and analyze large datasets through web and cloud technologies. Their goal is to stimulate new scientific applications and help researchers adapt to increasing data volumes. The platform allows scientists to share data access points, processing chains, and collaborate across distributed systems delivered as a service. Terradue is focusing on new services like data and software as a service to create marketplaces and leverage linked open data. They are also exploring how to use analytics and human resources like data scientists to help optimize the platform.
2. About us
• An Italian SME based in Rome, created in 2006 as a
European Space Agency (ESA) spin-off
• A UK subsidiary opened in 2011, based in the
Innovations Technology Access Center - Space Cluster
(Harwell Oxford, I-TAC)
• Experienced team: 11 members of 5 different nationalities!
most have worked for several years in the European
Space Agency
3. Our mantra : ‘digital earth scalers’
Develop innovative solutions that aim at
better prepared people for the new digital,
computational science
9. Markets & Services
Connect scientists to the Web and Cloud
resources that can empower their research
Forward
Horizon 2020 &
Copernicus
2008-2014
EC FP7 projects
2006
ESA G-POD
http://gpod.eo.esa.int/
The European
Earth Observation
Programme
Cloud Linked Data for
earth sciences
Grid-Processing on Demand
10. Why us ?
Stimulate new scientific
applications development
Integrate capacities in a multimission setting
Adapt to upcoming increasing
data volumes
Deliver new products for
decision-aid
Market on-demand services,
through one-stop-shops
Open Web Cloud Computing technologies can
innovate the way Earth Sciences researchers
work share.
11. A changing IT economy
Infrastructure / Platform / Software
Commodities provided as a Service
The economics of cloud computing are driving down the cost
structure of business so far and so fast that it s scary
Google CIO, May 2012
There is a potential community of ~50.000 independent
software vendors in the EU which require new ways to offer
solutions and services
EC Experts Group, March 2012
12. New services
Data as a Service
• Create Data Marketplaces
• Leverage Linked Open Data (LOD)
• Design for data quality traceability from the start
Processors (software) as a Service
• Sell usage right at low cost units (pay-as-you-go)
• Provide entry-level online presence
• Design for usage monitoring from the start
14. A platform for scientists
Share data access points
processing chains
Collaborate across
distributed Platforms
delivered as a service
Streamline web and
cloud APIs
Leverage computing
clusters
15. Concept of operation
Commodity Providers
- Elastic Cloud Computing
- Large storage
- High bandwidth network
DevOps Team
- Service Provisioning Monitoring
- Data Provisioning
- Virtual appliances maintenance
Partners
Users
- Application Engineering
- Software-as-a-Service
- Open Collaborative Developments
16. New development cycles
Python, CDAT, R, BEAM, BEAT
• Scientific developers use Cloud Sandboxes to !
implement, test and validate algorithms
• Includes tools for data discovery, parallelization, and
results validation
17. Cloud Compute deployments
• After validation, researchers deploy algorithms to a
selected Cloud Infrastructure
• Deployed Processor as a Service access complete
pools of datasets,!
run application !
intensive computing, !
and store processor !
outputs back to the lab
19. Platform engine: swarming
Open Science – A Future Vision
• Rapid deployment of data and processors
• Quality-driven aggregations of sensor feeds and earth models
• Scalable services providing cost-effectiveness
• Scientific information velocity
20. Platform gears: analytics
What does it take for a Cloud Sandbox developer to test a job chaining ?
What are the data needed for an area of interest at a given time span ?
What are the data archive gaps compared to the platform usages ? …
Analytics can answer such questions, from assets such as:
• Volume of information generated from the platform user logs,
sandbox trials, platform searches and tickets issuing, website
browsing patterns, login counts…
• Volume of information available from open data on the Web,
within OpenSearch catalogs, scientific blogs, twitter feeds,
stack overflow posts
• …
21. Platform gears: human resources
Know these jobs ? Come talk with us about how you see the future !
Data Scientist
- dig into data deluge resources
- assess opportunities with business management
- develop algorithms to find patterns and correlations
- link back insight to our DevOps team to prioritize sprints
Marketing Specialist
- help build customer dashboards in the
platform
- help management to make faster
strategic decisions
- communicate towards customer
prospects on our innovations