From space to spin-out – computing driven by astronomy in UCD
05 May 2015
Authors: Dr Sheila McBreen, lecturer, UCD School of Physics and Prof Lorraine Hanlon, associate professor, UCD School of Physics
Parameter Space Ltd is a spin-out company from the UCD Space Science group, founded by astrophysicists Sheila McBreen and Lorraine Hanlon in late 2014. The rationale to create the company came from the growing realisation that the techniques and instruments being developed in the group for astrophysics research have potentially a much wider realm of applicability.
Over the last decade, the group’s research has been mainly focused on the study of cosmic explosions called gamma-ray bursts (GRBs). These short-lived and unpredictable events happen about twice per day in the observable universe. Over several seconds, they produce more energy in gamma rays than our sun will produce at all wavelengths over its 10 billion-year lifetime.
To make progress solving the 50-year-old puzzle of GRBs required the concerted effort of the world’s most capable space and ground-based observatories. It is now widely believed that most GRBs are the signatures of the collapse of stars that are much more massive than our own sun. The penetrating power of gamma rays, combined with the extraordinary luminosity of the bursts, allows them to be used as beacons of star formation (and death) back in time to the very first population of stars that formed. A subset of even shorter bursts is most likely caused by the merger of two neutron stars, or a neutron star with a black hole.
Space Science Group
The Space Science Group in UCD has been involved in the European Space Agency (ESA) ‘INTEGRAL’ gamma-ray satellite since the scientific community submitted the initial concept to ESA in the 1990s. INTEGRAL was launched in 2002 and is still operating perfectly. The group’s major contribution (in collaboration with the Dublin Institute for Advanced Studies and IT Tallaght) was in the design and provision of the on-ground software analysis tools for the Optical Monitoring Camera instrument.
Before returning to Dublin in 2008, McBreen spent a number of years at the Max Planck Institut in Bavaria, Germany, as a member of the NASA Gamma-Ray Burst Monitor (GBM) team, currently operating on the Fermi satellite. McBreen’s role was in analysing real-time data, disseminating the results rapidly to the community and developing new software tools for the scientific exploitation of the GBM data.
The group is currently engaged in research into gamma-ray bursts, pulsars, terrestrial gamma flashes (which are gamma-ray events associated with thunderstorms) and a number of hardware projects to develop instruments and detectors for ESA.
One of the major challenges facing future astronomy missions is that the on-board instruments are so sensitive and have such large fields of view that the volume of data is increasing to the point that astronomers will have to change they how work. Traditionally, astronomers submit a proposal to an observing facility and, if accepted, they obtain observations of their target source. These data are then downloaded and analysed by the scientist who might request further observations at other wavelengths or download publicly available data from other sources and surveys.
Generally, the scientist works on individual sources, or groups of similar sources, and although the analysis can be very complex (such as for INTEGRAL or the Fermi Large Area Telescope) and may need to be processed on dedicated servers, the datasets are not so prohibitively large as to prevent users downloading products and working locally.
This is changing now that we are in the era of large-scale sky surveys. Ground- and space-based instruments with wide fields of view and/or scanning capabilities now observe the sky continuously and build-up all sky maps, opening up a huge discovery space. The European Space Agency mission, Gaia, is an example of one such observatory.
ESA’s Gaia mission
Gaia was launched in December 2013 and is currently orbiting the second Lagrange point (L2) at a distance of 1.5 million kilometers from Earth in the anti-sun direction, co-rotating with Earth in its one-year orbit around the sun. Gaia observes with two mirrors continuously and its primary objective is to measure the precise positions, motions and luminosities of one billion stars and to discover thousands of planets around other stars and supernovae.
To give just one example of the challenges this mission has posed, consider that during the five years of scientific measurements, the instantaneous orientation of the rotating Gaia instrument in space must be determined to an accuracy of about 100 microarcseconds. Since the spacecraft and optics have a diameter of three metres, this implies determining the relative location of any two parts on its outer periphery to about 1.5 nanometres with respect to inertial outer space!
Scientifically valuable information is contained in the data stream that results from the collection of photons in the approximately 100 on-board charge-coupled devices. However, the raw data is indecipherable, since Gaia picks up only fragments of the parameters it needs in each scan of its one billion+ targets. A large and complex data analysis must follow in order to piece together Gaia’s signal and translate it into the positional and physical parameters useful to scientists. To meet the extraordinary data analysis challenge imposed by the key science goals of the mission, more than 80 institutes, involving over 500 researchers around the world, formed the Gaia ‘Data Processing and Analysis Consortium’, which has worked for about eight years to solve these problems.
Data volumes in astronomy have now surpassed what is possible to visually inspect by even large teams of astronomers and volunteer citizen scientists. This necessitates an increasingly central role of software and hardware frameworks in the process of scientific discovery from astronomical archives, supplanting the traditional roles of humans in the loop. The way in which an astronomer engages with such homogeneous archives is therefore changing fundamentally.
Statistics take on particular importance, since an astronomer interested in a particular source may want to search for similar sources in a number of different surveys. It is undoubtedly a challenge to implement such advanced statistical capability for mission archives and must be automated.
Considering the much richer data environment of the early 2020s, when the final Gaia archive will be available over the internet, astronomers will want to connect with ground-based (e.g. the Square Kilometer Array) and other space-generated archives. As complex applications are developed, there will be a requirement to run a distributed application accessing one or more complete archives without pulling entire datasets over the internet.
Many applications already running on mobile phones or web browsers are just light front ends to functionality running on a remote server elsewhere. Virtualisation is also realised on the server end, where it is used for load balancing. This may also be needed on future archives depending on the weight of the application. Eventually everything must be in the browser or device, maintaining the appropriate interface, and what is running should probably not require any big download. This will require an application programming interface and development environment for creating the distributed applications.
Computing and astronomy
Parameter Space won a contract with ESA to develop a portal to the Gaia archive, which will allow advanced analysis tools to run close to the archive. The company will also develop a set of advanced tools to analyse transient sources (such as gamma-ray burst events) and work with two other companies – Space Systems Finland and Fork Research in Portugal – that are writing tools for temporal, spectral and visualisation analyses. Developers will also be able to submit their own analysis tools, which may also be publicly available. It is a challenging computing task driven by astronomy.
The Software Engineering Handbook of the European Co-operation for Space Standardisation (ECSS) is used to inform all aspects of the software development process – from requirements definition, through to management, implementation, validation, delivery, acceptance and operation.
Companies that can demonstrate successful engagement with ESA through software implementation according to these ECSS standards are in a strong position to gain added customers, who appreciate the rigour of the approach taken. Independent studies of Ireland’s membership of ESA have shown that engagement with ESA provides credibility and reputational benefits for the company.
With the implementation of the Gaia portal, more and diverse users will be motivated to take advantage of Gaia’s unique legacy, if the right tools are provided for them to enable their science or interest. Future-proofing the design, not only through the best use of current computing paradigms but also through consideration of ease-of-use, functionality, performance and scalability, offers in addition the potential for applicability to existing archives and future ESA science missions, such as Euclid.http://www.engineersjournal.ie/2015/05/05/space-spin-computing-driven-astronomy-ucd/http://www.engineersjournal.ie/wp-content/uploads/2015/04/Gaia-Spacecraft-1024x683.jpghttp://www.engineersjournal.ie/wp-content/uploads/2015/04/Gaia-Spacecraft-300x300.jpgTechsoftware,space,UCD