ADASSXIV Pasadena

Paper topics

Talks

MaxBCG: The Importance of Database Systems for the Virtual Observatory and the Grid

Authors María A. Nieto-Santisteban, Alexander S. Szalay, Jim Gray, Aniruddha R. Thakar, William J. O’Mullane, and James Annis

The Maximum-likelihood Brightest Cluster Galaxy (MaxBCG? ) application searches for galaxy clusters that span a large dynamic range of cluster masses and provides good redshift estimates using the cluster red sequence galaxies. MaxBCG? was developed originally in tcl using the SDSS Astrotools package and ran on the Terabyte Analysis Machine (TAM), a 10-CPU cluster specially tuned to solve this type of problem. The same application code was grid-enabled and used to test the Chimera Virtual Data System created by the GriPhyN project. As is common in astronomical file-based applications, the TAM and Chimera implementations used hundreds of thousands of files for the computations, files served from the SDSS Data Archive Server. The data required to run the MaxBCG? is also available in the SkyServer database. At Johns Hopkins University, we have implemented a version using SQL on a SQL Server 2000 cluster that processes the same sky area an order of magnitude faster than TAM. We are also working on an efficient grid-enabled system that does not require moving large volumes of data across the network. When the user submits the application, upon authentication and authorization, the SQL code is deployed and executed on the available data-grid nodes hosting the SkyServer database system.

In this paper we describe why and how the Grid and the Virtual Observatory can take advantage of Database Management Systems.

---

OpenSkyQuery & OpenSkyNode - the VO Framework to Federate Astronomy Archives

(demo in VO booth) (wil/tamas)

Authors William O'Mullane, Tamas Budavari, Vivek Haridas, Nolan Li, Tannu Malik, Masatoshi Oishi, Alexander Szalay, Aniruddha Thakar, Ramon Williamson

OpenSkyNode and ADQL are the major new steps in the Data Access layer of the Virtual observatory. Presented as a poster last year OpenSkyQuery(OSQ) now has a functioning portal which allows cross matches between catalogs on thirteen nodes. The portal also allows a user to upload their own list of source to be cross matched with the registered SkyNodes. This is one of the first systems utilizing of the IVOA's nascent standard Astronomical Data Query Language(ADQL). We shall demonstrate the functioning of the portal as it communicates with the nodes and the NVO searchable registry. We shall present the OSQ architecture and explain its use of WebServices as an open standard, Java and C# implementations of OpenSkyNode demonstrates that the open standard does indeed work. We shall also discuss the newest ADQL0.8 specification which will allow ADQL to be used to query registries in the VO indeed the the NVO searchable registry is already accepting ADQL0.7.4. and implements OpenSkyNode0? .7.4.

---

New Features for VO-Enabled Data Intensive Science with the SDSS Data Release 3

Authors A. Thakar, A.Szalay, J.Gray, W. O'Mullane, T. Budavari, M. Nieto-Santisteban, G. Fekete, N. Li, R. Lupton

With the third data release (DR3), the SDSS Catalog Archive Server (CAS) serves up 3 Terabytes of catalog data for nearly 200 million celestial objects via the SkyServer web pages at skyserver.sdss.org. Several changes and new features have been added that facilitate data-intensive science with the data and advance VO standards and technologies.

1. The CasJobs batch query system unveiled at ADASS XIII now allows stored procedures and functions in the user's MyDB? . This enables users to bring their program to the data rather than the other way around, which is far more efficient for data-intensive applications. We replicate user account details between installations to allow distributed CasJobs configurations. This form of replication raises issues that are crucial to the VO community. We are also attempting to secure the system using X509 certificates through WS-Security. This ties in with the distributed storage effort within the VO.
2. New color images in the Visual Tools based on the ASINH stretch that are deeper and more feature-rich. We briefly discuss the algorithm.
3. The traffic logging system is now quite detailed and extensive, allowing many useful and interesting statistics to be gathered. The Support Services specification proposed for IVOA includes a logging section modeled on this apporach.
4. A much faster HTM (spatial index) library with an upgraded interface.
5. An enhanced object crossid facility allowing users to run their own SQL query on the matching objects.

We briefly describe these features and list future enhancements anticipated with SQLServer Yukon and SDSS-DR4.

---

Astronomical Computing within Data Archives

Authors Adrian Pope, Alex Szalay, and Jim Gray

As astronomical datasets continue to grow in size we begin to face serious issues with transporting data from archives to computing resources used to do calculations with the data. We have begun to attack this problem by performing as many calculations as possible within the archive. One example is the use of a SQL Server archive with Hierarchical Triangular Mesh (HTM) spatial indexing software to pixelize the low redshift Sloan Digital Sky Survey (SDSS) Main Galaxy Sample (MGS) before performing Large Scale Structure (LSS) analyses. The resulting counts-in-cells data is an order of magnitude smaller than the input catalog, dramatically reducing the amount of data transported to external programs. We also present a system for analytically calculating the spatial relationships between spherical polygons within an archive that has been used to analyze the complex geometry of SDSS survey footprint and masks, a necessity for preparing samples of data suitable for detailed LSS studies. The software was written (by Alex Szalay and Jim Gray) in the high level Transact-SQL language, with the use of the HTM library. Finally we describe the creation and utilization of a Monte Carlo realization of the SDSS for studying the effects of calibration errors and masking on LSS studies. The realization contains hundreds of millions of random points and was created using a cluster of SQL Server machines.

---

hardware and performance, bulk loading (jan) T

---

Posters

FITS Image Cutout (wil/vivek) P

Authors Vivek Haridas, William O'Mullane, Alex Szalay, Tamas Budavari, Maria Nieto-Santisteban, Samuel Carliles, Aniruddha Thakar

The Sloan Digital Sky Survey is on its way to build a very large map of the universe. We look at implementing web services that help in getting image cutouts from the FITS files in the SDSS archive. These cutouts would be provided in the traditional FITS format as well as in the gray scale jpeg image format.

Some amount of complexities exists in creating cutouts in both the formats. The process of getting cutouts in the FITS format would require World Coordinate System values to be recalculated for the newly created FITS files according to the positioning of the cutouts in the original archive images. Also, there is a possibility of the cutouts not existing completely in any single FITS file in the archive.

The jpeg image cutouts provided are actually a gray scale visual rendering of the FITS Image cutouts. In addition to the above complexities, sky noise calibration, arcsinh scaling and gamma corrections are performed on the image data.

K-12 Education with the National Virtual Observatory (jordan/wil) P

Authors Jordan Raddick, Wil O'Mullane

We have created the first K-12 education activity using data provided by the National Virtual Observatory (NVO). The activity uses NVO’s “Datascope,” which delivers multiwavelength data on single objects requested by the user.

The activity, “Adopt an Object,” was suggested by Heidi Kaiter, a middle school science teacher from Concord, MA. It is designed for middle school students but could be adapted for high school and Astro 101 students as well. Each group of students selects a well-known sky object (star, galaxy, or nebula) to “adopt” for detailed study. Students use the datascope to look up multiwavelength images, observations, and catalog data for their object. The activity ends with a brief oral presentation and written report summarizing what each group has learned about their object.

The activity includes a complete teacher’s guide, with a lesson plan, additional resources, and correlations to national education standards. We are currently looking for K-12 teachers to field test the activity in their classes. We are also developing several other activities for K-12 and college teachers using NVO data.