Virtual Research Environment

The polar sea ice cover is a rapidly changing terrain, mapped regularly by several EO missions, for example, NASA’s Aqua/Terra MODIS, ESA’s Sentinel series as well as other third party missions.

With a rapid reduction in the extent of sea ice in the Arctic, new shipping lanes are opening up and maritime activities are increasing along new routes. The majority of ships wanting to use these routes need to spend an average $50,000 a day to use the services of ice-breaker escorts (Wall Street Journal, 23^rd Sep 2013). The high price and infrastructure poses a prohibitive barrier to small ships, research groups and indigenous population to navigate in this region.

A preliminary solution presented by CGI at the NASA London Space Apps Challenge 2016 called “ICETRACK” advocated an innovative way to address this problem by using daily Arctic ice cover data from the MODIS platform published by NASA GIBS archive and applying a simplified version of the A* algorithm similar to path planning on Mars rovers. The key theme of the solution was to create a service for research oriented vessels, small ship operators and local navigators to find optimal routes through broken ice sheets which open/close new channels on a daily basis.

A prototype was implemented which displays sea-ice data from EO scenes and allows the end-users to assign waypoints on the map, to be used as starting points for the path planning algorithm which can take the map with the waypoints and draw the optimal path between the waypoints.

This approach is being complemented with a combination of Sentinel-1 and other multi-mission data for frequent updates of floating ice movements and incorporating predictive analysis based on previous ice movements.  ICETRACK intends to create a vastly cheaper and safer solution for local navigation in arctic region minimizing the need for ice-breaking but optimizing exploitation of natural channels.

© 2016 CGI GROUP INC

FabSpace 2.0 aims at making universities open innovation centres for their region and improving their contribution to the performance of societies. It will concentrate on one innovation area with high expected socio-economic and environmental impact: geodata-driven innovation, by leveraging space data in particular.

When Universities must endorse a new role as co-creators of innovations in the context of Science 2.0 principles, the future realisation of open data as a new innovation scene needs to set up a creative environment in which developers from the civil society or industry or the academic research, public administrations and civil organisations can meet, work together, co-create new tools and business models.

The proposal consists in a new kind of fab labs: FabSpaces. A fab lab is a one-stop shop access to any materials, machines and tools to digitally manufacture new products, a FabSpace will be a one-stop shop-access to a range of data in particular space data, free software & data processing tools, to develop new applications.

FabSpace 2.0 will provide a new free-access service and place dedicated to collaborative data-driven innovation in 6 European Universities (13 at the end of the project). Through online support, entrepreneurship & innovation leadership trainings for students and researchers, the human capital among FabSpace users will be enhanced. The link between Universities, industry, the public sector and civil society will be strengthened with 18 local and 3 European contests consisting in developing new applications to tackle challenges launched by ONGs & companies. 1500 students and researchers are expected to use FabSpaces, half of them will participate in contests and 21 teams will be awarded a grant to be used for product development only. FabSpace sustainability will be ensured before the end of the project.

By doing so, FabSpace 2.0 intends to make universities regional catalysts and leaders of innovation in the area of data-driven innovation.

As well Fasbspace 2.0 will contribute to the knowledge of student from different origins, technical and non technical, of the potentialities of space data.

Two key factors in delivering an effective platform for Open Science are data discoverability (i.e. finding data that are of interest) and high-throughput data access. At the Centre for Environmental Analysis (CEDA), we currently hold approximately 3PB of data in our archive, and with data streams from missions such as the Sentinels coming online and the upcoming CMIP6, the volume of data in the archive is increasing at an unprecedented rate. With such a rapid increase in the volume of available data, it is becoming increasingly difficult to present that data to users in a way that is both relevant to them and performant. By providing a private cloud co-located with the CEDA archive, JASMIN is uniquely positioned to enable the development of innovative data-intensive services by allowing developers to deploy their own software in an environment where it can benefit from high-performance access to data in the CEDA archive.

In order to increase the discoverability of data, CEDA are indexing the entire data archive of earth observation and environmental science datasets into ElasticSearch.  Initial results are demonstrating the potential of this technology with data indexed by named scientific terms defined in controlled vocabularies.  In this way powerful faceted search interfaces can be developed enabling users to quickly find datasets that meet particular criteria. The second aspect, high performance access to data, is being approached by re-orientating JASMIN’s network topology for its data-access services (including FTP, HTTP, OPeNDAP) for high throughput. These services will sit in a dedicated portion of the network that is optimised for large bandwidth (a data-transfer-zone, or DTZ), and will use a combination of high-performance software and load-balancing to achieve optimum performance. These services will be visible to the internet, but services running in the JASMIN cloud will see an additional benefit from a high-bandwidth link to the DTZ.

A number of projects, such as the ESA CCI Open Data Portal, have already demonstrated the value in having a private cloud co-located with the archive by developing web portals to present data stored in the CEDA archive in a user-friendly manner. The OPTIRAD project has also demonstrated the potential of JupyterHub as a collaborative research environment which, when combined with high-performance access to archive data, can be an extremely powerful tool. By providing supported services for discovering and accessing data from the archive in a performant manner, CEDA aims to increase the performance of existing services and facilitate the development of innovative, data-intensive services in the future.

The Earth Observation Data Centre for Water Resources Monitoring (EODC) was founded in 2014 to enable the exploitation of Big Earth Observation (EO) data and to facilitate the cooperation among partners from science, public and the private sector. Through its IT infrastructure components, the EODC facilitates collaborative method and software development and testing, as well as archiving, processing and distribution of EO data.

Since day one, EODC has been downloading and ingesting Sentinel-1 and Sentinel-2 data from the European Space Agency’s (ESA) data hubs, with the aim to provide a complete and up-to-date archive to its partners. In parallel, EODC’s private cloud infrastructure has been extended and evolved to meet the computing requirements of an increasingly demanding EO community. During the last 2 years of EODC’s existence, a number of mutual projects have been carried out by EODC partners, each one resulting in individual components, data or information products further extending EODC’s capabilities.

After this initial build up phase, the EODC cooperation network now faces the challenge to integrate these individual results to create operational EO services that can deliver added value to EODC partners and to the whole EO community. With this contribution, we introduce recent co-operative developments achieved by selected EODC partners, grouped in (1) infrastructure and (2) data and information services.

As for infrastructure, we provide updates on (a) the EODC’s redundant data acquisition sources, (b) the EODC data archive concerning Sentinel and other EO data sets (e.g. Landasat), (c) our processing framework and compute clusters, intended to handle batch computation of all sorts of EO data and information products available on the EODC archive. We also discuss (d) recent developments as regards the federation of infrastructure and data catalogue services with cooperating data centres.

Based on these advances, we will showcase some of the data and information services that have been built on top of the EODC’s infrastructure capacities. Covering a wide range of application fields, this includes Sentinel-1 Level 1 data composites, Sentinel-2 Level 2A on-demand product generation, Sentinel-1 wetland mapping and classification, as well as Sentinel-2 crop status monitoring.

The rise in cloud computing as an e-infrastructure model is one that has the power to democratise access to computational and data resources throughout the research communities. We have seen the difference that Infrastructure as a Service (IaaS) has made for different communities and are now only beginning to understand what different models further up the stack can make. It is also becoming clear that with the increase in research data volumes, the number of sources and the possibility of
utilising data from different regulatory regimes that a different model of how
analysis is performed on the data is possible. Utilising a "Desktop as a
Service" model, with community focused applications installed on a common and
well understood virtual system image that is directly connected to community
relevant data allows the researcher to no longer have to consider moving data
but only the final analysed results. This massively simplifies both the user
model and the data and resource owner model. We will consider the specific
example of the Environmental Ecomics Synthesis Cloud and how it could easily be
generalised to other areasKeynote Speaker