The TVO Archive for Cosmological Simulations: Web Services and Architecture

Computing resources such as Grid or Beowulf clusters are improving the possibility of performing simulations on a cos- mological scale. The box size and the number of simulated particles are increasing, giving us helpful theoretical results tools to analyze the development and the evolution of a cosmo- logical system based on a particular pattern. A cosmological simulation is an expensive result in terms of CPU time and often in terms of human effort: this result is priceless and must be exploited as much as possible. Moreover, a cosmological simulation archive can include an extremely large amount of data; this implies strong difficulties in analyzing and moving these kind of data, and so it is impor- tant to develop a system that analyzes the cosmological data where they are stored. We have built a software layer that can easily allow us to handle data. The challenge is to develop and supply to the community a set of services for data handling by providing a user-friendly access to a huge amount of hetero- geneous data and also an optimized way to process and analyze these data in a distributed environment. We can identify three classes of users interested in a data set resulting from a cosmo- logical simulation: the group that performed that simulation; the whole astrophysical community; the nonscientist community, if we consider that a simulation can be used as an auxiliary tool in several education activities. The Web services and the database access is provided through two TVO Web portals. Such an infrastructure helps avoid the direct execution of several Web services and provides an effective interface to the database (DB) service

PUBLICATIONS OF THE ASTRONOMICAL SOCIETY OF THE PACIFIC, 120:933–944, 2008 July © 2008 The Astronomical Society of the Pacific All rights reserved Printed in U.S.A The TVO Archive for Cosmological Simulations: Web Services and Architecture A COSTA,1 P MANZATO,2 U BECCIANI,1 M COMPARATO,1 V COSTA,1 F GASPARO,2 C GHELLER,3 A GRILLO,1 M MOLINARO,2 F PASIAN,2 AND G TAFFONI2 Received 2007 December 21; accepted 2008 June 24; published 2008 July 29 ABSTRACT In order to offer an intuitive but effective access to a growing number of cosmological simulations, we have developed the Italian Theoretical Virtual Observatory project (ITVO), as described by Pasian and colleagues in 2006 In this work we describe two Web portals as two ways to access and share complex data coming from numerical astrophysical simulations We present a set of Web services aimed at offering services such as Simple Numeric Access (ProtocolSNAPSimple Numeric Access Protocol), as described by Gheller and colleagues in 2006, and Randomizers dealing with different data formats The Web services technology allows us to run a particular task (a SNAP job, for instance) close to its data, avoiding an expensive data transfer Online material: color figures INTRODUCTION services and provides an effective interface to the database (DB) service Computing resources such as Grid or Beowulf clusters are improving the possibility of performing simulations on a cosmological scale The box size and the number of simulated particles are increasing, giving us helpful theoretical results tools to analyze the development and the evolution of a cosmological system based on a particular pattern A cosmological simulation is an expensive result in terms of CPU time and often in terms of human effort: this result is priceless and must be exploited as much as possible Moreover, a cosmological simulation archive can include an extremely large amount of data; this implies strong difficulties in analyzing and moving these kind of data, and so it is important to develop a system that analyzes the cosmological data where they are stored We have built a software layer that can easily allow us to handle data The challenge is to develop and supply to the community a set of services for data handling by providing a user-friendly access to a huge amount of heterogeneous data and also an optimized way to process and analyze these data in a distributed environment We can identify three classes of users interested in a data set resulting from a cosmological simulation: the group that performed that simulation; the whole astrophysical community; the nonscientist community, if we consider that a simulation can be used as an auxiliary tool in several education activities The Web services and the database access is provided through two TVO Web portals Such an infrastructure helps avoid the direct execution of several Web COSMOLOGICAL SIMULATIONS Cosmological simulations incorporate a range of physics phenomena: gravitation, gas dynamics (adiabatic or with radiative cooling and heating), radiative transfer, and chemistry These computations are almost always performed using comoving spatial coordinates and periodic boundary conditions so that a finite, expanding volume is embedded in an appropriately perturbed background spacetime More in detail, the SPH (SmoothParticle Hydrodynamics) is a Lagrangian (particle-tracking) method for integrating the fluid equations invented by Lucy 1977 and Gingold & Monaghan 1977 The fluid variables (baryon density, velocity, temperature, for example) are followed using particles of fixed mass representing fluid elements The method, an extension of N-body methods, makes it relatively easy to add SPH to existing cosmological simulation codes Initial conditions for simulations of structure formation consist in specifying the background cosmological model and the perturbations imposed on this background The cold dark matter (CDM) model is the platform on which simulations of cosmic structure formation matured into a powerful theoretical result The CDM model adopts parameter values H and Ωc ¼ Ωb , where Ωc and Ωb give the present mean mass density of CDM and baryons, normalized to the critical density 8πG=3H 20 , respectively The first gravitational N-body simulation of interacting galaxies was performed using an analog optical computer; see Holmberg (1941); nowadays many types of simulation exist, including N-body, Mesh, adaptive mesh refinement (AMR), N-body+SPH, N-body+Mesh, and Nbody+AMR, and are carried out using high performance computing resources or the Grid infrastructure The aim of these INAF-Osservatorio Astrofisico di Catania, Italy; alessandro.costa@oact.inaf it INAF-SI / Osservatorio Astronomico di Trieste, Italy; manzato@oats.inaf.it CINECA—Casalecchio di Reno, Italy 933 934 COSTA ET AL methods is to simulate many different kinds of objects, for example: dark matter haloes or globular clusters Numerical simulations of galaxy clusters have been used among others for the following goals: Understanding the general processes of formation and evolution of single clusters (galaxy evolution and dynamics, intracluster medium); Testing observational methods of mass estimation (for both dark matter and baryons); Using the distribution of X-ray temperature (or luminosity, mass, cluster velocity dispersion) and its evolution to constrain cosmological parameters; Using substructure, morphology, shape, or radial profile to constrain cosmological parameters N-body simulations have allowed cosmologists to address a straightforward question: what is the shape of dark matter halos formed by hierarchical clustering? This question is relevant observationally to the cusps of elliptical galaxies (e.g., Faber et al 1997) and the profiles of clusters of galaxies Therefore, to answer this question, we are making data produced by simulation available to the whole community; to this we have organized a Theoretical Data Archive to make simulated data easily available under the VO standard The ITVO is a pilot project (see Pasian et al 2006) in which we create a distribution archive and we use Grid technologies; see Taffoni et al (2006), to access data that may be not located in the same system but in different locations As a test bed, we start to deal with three different simulations The first one is a large cosmological hydrodynamic simulation (see Borgani et al 2004), which used the massively parallel tree N-body+SPH code GADGET-2.0 (see Springel et al 2001), to simulate a concordance Λ CDM cosmological model within a box of 192 h Mpc The cosmological parameters assumed were Ωm ¼ 0:3, Ωb ¼ 0:04, H ¼ 70 kms Mpc , and σ8 ¼ 0:8 This simulation regarded star formation, radiative cooling, metal production, and galactic wind It produced 102 snapshots for a total amount of approximately 1.2 TB of raw data All these parameters and results have been included in the level of the DB Subsequently, from the raw data of the initial big box, a first level of postprocessing has been implemented to identify the groups and clusters of galaxies, first applying a friend-of-friend (FOF) halo finder to the distribution of dark matter particles, then running a spherical overdensity algorithm: 400 haloes and 72 clusters were identified The total number of files stored in the second level of the DB consists of 1986 boxes that refer to the clusters at different redshifts Then, another step of postprocessing has been implemented on the clusters at red shifts equal to zero: these are 117 clusters for which we can extract 2D maps for the density, spectroscopic-like temperature, mass-weighed temperature, SunyaevZel’dovich emission, and X luminosity in the 0:5 × 1044 erg s Þ band, the corresponding profiles for the density, spectroscopic-like temperature, mass-weighed temperature, and emission-weighed temperature These more refined data are stored in the third level of the DB The second simulation is an AMR (Norman & Bryan 1999), grid-based hybrid code (N-Body+hydrodynamic), designed to simulate the cosmological structure formation; it used the Enzo code.4 We stored two simulations made with this code, both of them simulate a Lambda CDM universe with the following cosmological parameters: Ωm ¼ 0:27, Ωb ¼ 0:044, H ¼ 71 kms Mpc , and σ8 ¼ 0:94 The third simulation consists of a set of N-body results that were obtained using the FLY code (see Becciani et al 2003); typically the N-body simulations were performed on boxes of sizes between and 120 h Mpc, with the number of particles varying between 1283 and 4003 , and for different CDM cosmological models, ranging from ΩCDM ¼ and ΩΛ ¼ to the most popular “concordance” model (ΩCDM ¼ 0:3 and ΩΛ ¼ 0:7) WEB PORTAL FACILITIES The static part of the ITVO Web portals is realized in HTML with CSS, while the dynamic one is written in PHP,5 a server-side scripting language that is especially suited for Web development and can be embedded into HTML; see Costa et al (2006) The Web server used for our purposes is Apache 2.0,6 the most famous open-source HTTP server for modern operating systems 3.1 ITVO Portal in Trieste At this time, the database access is allowed from three separate levels of the Web interface accessible from the Italian Astronomical Archives Center (IA2) Web site.7 The metadata are stored in an Oracle 10g database The level interface allows you to find raw data boxes of the simulation, the next level (level 1) to find and download data of the simulated clusters at several red shift values, and the last one (level 2) allows you to search and download the maps of the galaxy clusters, to visualize and print the graphics of many quantities dynamically generated to save a VOTable or FITS BINTABLE of these quantities, to visualize the header and the preview of the maps of the FITS file (also with a link to the Aladin tool applet) and to save, as a VOTable, the catalog obtained as a result of the query to the DB In the Web interface, it is also possible to personalize the SQL query (see Fig 1) and its results Some examples of scientific query for every level of DB are The Enzo Code is available online from the University of California, San Diego, http://www.cosmos.ucsd.edu/enzo/ Hypertext Preprocessor (PHP) is available online at http://www.php.net/ The Apache Software Foundation is found at http://www.apache.org/ More information about IA2 is available at the URL http://wwwas.oats.inaf it/IA2/ITVO/ 2008 PASP, 120:933–944 TVO ARCHIVE FOR COSMOLOGICAL SIMULATIONS 935 FIG 1.—ITVO@Trieste Web portal: query form, personalized SQL form, and graphics of theoretical quantities performed on the fly See the electronic edition of the PASP for a color version of this figure 1.Discover a snapshot in an assigned red shift range and/or in a specific cosmological parameter (Ωm , ΩΛ , H , for example) and/or originating from a specific algorithm; 2.Find all galaxy clusters in a given red-shift range and/or within a specific number of a kind of particles and/or in a specific boxsize range; 3.Find all maps of galaxy clusters in a selected mass-weighed temperature range and/or a virial radius range and/or X luminosity and/or according to the ratio of two selected quantities This functionality permits a search over a large amount of scientific parameters, allowing a simple and direct query to find 2008 PASP, 120:933–944 the data that better satisfy the characteristics of a research typology 3.2 ITVO Portal in Catania The ITVO portal in Catania can be reached at at a separate Web site from the Trieste Portal.8 The database engine that contains the ITVO portal simulations metadata is More information about the VO Portal is found at the URL http://itvo.oact inaf.it/ 936 COSTA ET AL FIG 2.—ITVO@Catania Web portal: query form and SQL editor See the electronic edition of the PASP for a color version of this figure PostgreSQL 8.1,9 an advanced open-source database The access to the simulations data takes place by an authentication system, based on user name and password A new user requests an account on the ITVO portal by clicking on “Register”: a registration form is shown and has to be filled out with user data After the submission of this form, if the user data has been correctly filled out, a confirmation email is sent to the user, and then the account is activated by the Web site administrator Moreover, it has a password-retrieving system; in fact, by clicking the “Lost Password” link, the “Password-Retrieving Form” (asking for the user name) is shown After the submission, an email will be sent to the user to inform him about the password-retrieving procedure The “ITVO Level 0” link brings up a form that allows the user to generate a query according to several parameters First, the user can choose the fields to be included into the query output by checking the left-side corresponding check boxes Then, the user can put restrictions to each of the simulations parameters by checking the right-side corresponding check boxes and by filling the corresponding fields By clicking the The PostgreSQL software is available at http://www.postgresql.org/ “SEARCH” button, the customized query will be generated and directly submitted to the database engine, while by clicking the “GO” button a text area containing the generated SQL query will be shown, allowing the user to edit and submit it (Fig 2) The query result will be a table containing the requested data Each record of this table contains the requested parameters and, on the left side, a radio button for its selection After the record selection, the user can choose some actions: download the binary simulation data, preview a picture of the binary data, download a snapshot of the binary data, or view and download the VOTable—an XML representation of the metadata (Fig 3); finally the “Go to Level 1” button will bring up the form for the extraction of a subsample of the binary data (Fig 4) DISTRIBUTED DATA ARCHIVE The DB has been designed to store all kinds of cosmological simulations Due to the variety of data to be stored, we decided to use a relational DB, whose query engines allow the user to make complex queries in a standard language, SQL, which is also the standard query language used by many tools developed under the IVOA standards (see Hanisch & Quinn 2003) The ITVO DB (see Manzato et al 2007) is a multilevel database, 2008 PASP, 120:933–944 TVO ARCHIVE FOR COSMOLOGICAL SIMULATIONS 937 FIG 3.—ITVO@Catania Web portal: query results and VOTable generator See the electronic edition of the PASP for a color version of this figure as can be seen in Figure Every level holds the data from one step of the whole data process Another similar structure was made by Lemson (see Lemson et al 2006) for the Millennium simulation that produced halos and galaxies and contains only the metadata of postprocessing products of the simulation Level of the database includes the description of the FIG 4.—Spherical SNAP form See the electronic edition of the PASP for a color version of this figure 2008 PASP, 120:933–944 algorithm, the computational and cosmological parameters used to make the computational run, the species of matter inserted in the simulation, the physical quantities linked to the particles or grid points, along with the format (HDF5 and GADGET at the moment in our case), resolution, and redshift of the output file The next level includes the link to the code used for the postprocessing, for example, the code used to extract the clusters and group of galaxies from the initial box and the metadata of these new output file, such as virial radius, virial mass, and temperature The last level refers to a more refined postprocessing: it stores all the metadata of the FITS files concerning the two-dimensional maps All these data are stored in tables that are linked to each other by foreign keys on primary keys to avoid any duplication of the information at every level There are also two auxiliary tables, one to take into account the units used for the different quantities and the other to take into account the rank value of the quantities Moreover, a theory Data Model (see Lemson et al 2007) is being developped by a group of interest within IVOA to define a model for the metadata to be associated with the publication of numerical simulations The theory Data Model consists of 938 COSTA ET AL FIG 5.—ITVO multilevel Database Structure: level 0, level 1, and level See the electronic edition of the PASP for a color version of this figure 2008 PASP, 120:933–944 TVO ARCHIVE FOR COSMOLOGICAL SIMULATIONS a UML diagram of classes including all the variables and quantities used and calculated in these simulations The data model has to be validated from our DB schema and another DBs contains the metadata of other simulations to enable future queries on distributed databases and archives WEB SERVICES A great effort has been required in order to deal with different data formats hosted by the archive The target is to create a sole interface for every service independent of the different data formats Following the suggestions of the IVOA theoretical interest group we can define a “snap-cutout request” (see Gheller et al 2006) as a service that forwards to the server the selection parameters in order to obtain the access to a subset of the raw data of a simulation This operation provides the following functionalities: extraction of a subset of data properly selected (data cutout), storage of the associated metadata in a VOTable delivered to the client, staging of the extracted data, and and their delivery to the client via http, ftp, or some other method In Catania we have implemented a “snap-cutout request” Web service that allows the user to obtain either a rectangular or a spherical subregion of the entire computational volume In order to set a rectangular area of the computational box and to obtain a data cutout you have to select the center of the subregion and the length of each of its sides For the spherical counterpart the center and the radius of the sphere are required Figure shows a spherical set-snap form in the ITVO portal Another useful Web service available in Catania is the “randomizer” that provides a random sample of the data set For this service you have to specify the final number of particles to be sampled This service is designed for data sets coming from particle-based data files such as N-body simulations and SPH The user can obtain a VOTable containing the metadata associated to the object of interest The staging of the extracted data is given to the client asynchronously via an ftp server The user can exploit these Web services by deploying its application in java or C++ The client application can run on the user platform invoking the Web services that will act on the remote dataset A great help is given by the Apache tool wsdl2java that can automatically generate the client classes both for java and C++ The Web services invoked in the remote system are written in java using the SOAP engine Axis.10 SOAP11 is an XML-based communication protocol, an encoding format for interapplication communication The Web services are coupled with a group of applications that deal with the remote data performing the requested operations This set of application in written in C and C++ in order to optimize the performances of each operation Using the PHP language we have built a section of the ITVO portal that can exploit the services provided by the ITVO so that the portal can act as a Web service client In Trieste a Web service has been implemented for the visualization of the 2D postprocessing data, as described in the next section As a second service, we are planning to implement a service that creates on demand a FITS BINTABLE and/or VOTable of 10 quantities of the galaxy clusters of the GADGET simulation from which the profile graphics with the TOPCAT tool can be easily extracted In the near future in Trieste we will make available a SNAP Web service for the GADGET and HDF5 Enzo simulation files WEB SERVICE ARCHITECTURE We describe here the ITVO services using two different diagrams: a Use Case Diagram and a Block Diagram The Use Case diagram describes only those features visible and meaningful to the actors who use the system (Fig 6) An actor is a role that an external entity plays in relation to the system When a Use Case delegates to another, the dependency is drawn as a dashed arrow from the “using” to the “used” Use Case and labeled with the include stereotype notation This means that the execution of the “using” (or calling) Use Case will include or incorporate the functionality of the “used” one SetSnap allows you to specify a snap-cutout request SetRand allows you to specify the number of particles in the output file for a randomized request GiveResult gives result to the Client Application StageData stages data on the server to be later retrieved by the client The data is only staged for a period of time and is eventually deleted by the service Figure shows a block diagram representation of ITVO Web services The dashed box is a system that contains Web services, applications, and the Simulation Data Archive: a particular job will run inside this schema The steps are (1) Authentication: the Web service requires authentication in order to accept each client request The authentication is based on a plain text file (MD5 algorithm with a 10 The Apache Axis Web Services can be accessed at http://ws.apache.org/ axis/ 11 Simple Object Access Protocol, http://www.w3.org/TR/2000/ NOTE‑SOAP‑20000508/ 2008 PASP, 120:933–944 939 FIG 6.—Use Case UML Diagram describing the Web services 940 COSTA ET AL posal is a process by steps: the first is in giving the possibility of finding all the kind of data stored under the IVOA standards including theoretical data The second, after selecting the data type of interest, is in returning a list of parameters to be filled in depending on the choice made before The user then will make a request selecting those fields The third step is in getting back the response by the server in a standard format, like VOTable, with the URL link to the data files or to the data services ASTRONOMICAL TOOLS FIG 7.—Block Diagram describing the Web services See the electronic edition of the PASP for a color version of this figure In the following section we will describe some useful astronomical tools that are fully compliant to the Virtual Observatory standards, giving the user an effective and interactive access to the data 8.1 VisIVO 128-bit hash value) (2) Logging: we have implemented a server-side multilevel log writer (debugging/production) It provides a log-level configuration and can keep trace of events and errors (3) Message Parser: this component parses the syntax of each request (4) Application Handler: this component creates a native process and gives information about it providing methods for controlling the process, waiting for the process to be completed, checking the exit status, and destroying the process Figure shows a spherical cutout of a cosmological N-Body simulation (FLY code) The original simulation has 16,777,216 particles, and the totally extracted particles are 79,936 This result has been achieved using the cutout Web service SetSnap and the ITVO portal as client We have visualized the two data sets using VisIVO At the moment the Web services are defined for GADGET and FLY files VisIVO (see Comparato et al 2007), is a data exploration tool oriented to astrophysical problems It provides a powerful tool for huge data analysis and visualization from both catalog queries and numerical simulations The software is specifically designed to deal with multidimensional data, and several physical quantities can be explored at the same time VisIVO blends high performance multidimensional visualization techniques and up-to-date technologies, with the interoperability concept to cooperate with other applications and to access remotely distributed data archives (Fig 9) VisIVO is integrated in the VO framework and is supported in the VO-TECH project The European Virtual Observatory - VO Technology Centre (VO-TECH) is a Design Study implemented as Specific Support Action funded by EU in the FP6 Program The VO-TECH project specifically aims at feasibility studies and design works REGISTRY SERVICE Inside the IVOA we are collaborating to find the right manner to include in the registry the publication of the simulated data archive and the services linked to these kinds of data The pro- FIG 8.—Spherical cutout obtained through the ITVO Web services using the portal as a client See the electronic edition of the PASP for a color version of this figure FIG 9.—VisIVO Data Exploration: Input Data Formats, Level Of Detail, and VOTable visualization See the electronic edition of the PASP for a color version of this figure 2008 PASP, 120:933–944 TVO ARCHIVE FOR COSMOLOGICAL SIMULATIONS 941 FIG 10.—Aladin tool Background panel: simulated galaxy clusters (top) and A85 galaxy cluster images by Chandra and XMM (bottom) See the electronic edition of the PASP for a color version of this figure 2008 PASP, 120:933–944 942 COSTA ET AL FIG 11.—Aladin tool loads theoretical data using an Apache-Tomcat Web service See the electronic edition of the PASP for a color version of this figure to integrate such new technologies into the EuroVO (see Becciani et al 2006b) In particular, it aims at intelligent resource discovery (ontology and the semantic web), data mining, and visualisation capabilities VisIVO (see A Costa et al 2006a), is a C++ tool running on both Windows and Linux, the Mac porting being in progress VisIVO supports different file formats: Binary files, VOTables, FITS, HDF5, ASCII, and the native data format VisIVO can deal with both structured and unstructured data and the internal data representation is a data table Structured mesh-based data not have a default graphical representation VisIVO can visualize them using volume rendering and isosurfaces Each column of the internal table data, in an unstructured data set, is a field that can be selected for the representation in the rendering window: the selection of three fields represents a view of data For example: the x, y, z coordinates of the particles from an N-Body simulation, or the R.A (Right Ascension)— Decl (Declination)—Vmag (Johnson magnitude V) fields for data from stellar catalogs In addition to their geometric position, points can be used to display further quantities, using colors and glyphs (3D shapes, such as spheres or cubes) If the data size is too large to be managed in memory, VisIVO allows the user also to extract a random subset of points It is also possible to select the points that lie in a region: the selection can be accomplished using either a rectangular sampler or the cluster finder utilities An important feature, the level of detail (LOD), has been recently included in the tool; it decreases the complexity of the displayed 3D object (structured and unstructured) allowing the user to move fast even if huge data are loaded—up to 16 million positions and velocities of an N-Body simulation with GByte memory VisIVO can be used for the visualisation of a “randomized” data set coming from the randomizer Web service of the ITVO This tool can help identify of a particular region in the computational box, and then the user can obtain all the data belonging to the numerical simulation in a spherical or rectangular region through the GetSnap service of the ITVO code, we added a new feature to the ALADIN tool to enable the search of simulated galaxy clusters by virial mass, virial radius, X luminosity, and mass-weighed temperature within the virial radius to be immediately compared with an observational one (Fig 10) The Aladin tool (Bonnarel et al 2000) is downloadable at the CDS URL.12 To enable the new features, a user also has to download a txt file.13 To add the ITVO menu and launch the Aladin.jar file, use the command: java -jar Aladin.jar -glufile=ITVO.txt This IVOA tool leans on a Web service that enables the possibility of making a distributed remote query to multiple databases The architecture of this service is based on a call to a servlet that holds the information about the selected parameters and queries the database giving as result an XML file, VOTable, that Aladin is able to ingest and open for visualization (Fig 11) 8.3 TOPCAT From the ITVO, at Trieste Web portal it is possible to launch a program to dynamically create a VOTable or FITS BINTABLE of 10 quantities of the simulated galaxy clusters and open it with the TOPCAT (Taylor 2005) tool to easily create the plots of these profiles, as shown in Figure 12 All of these VO standard tools can be connected to each other using the PLASTIC hub (Taylor et al 2006), a software designed for interoperability between astronomical VO applications The motivation for PLASTIC is the desire to leverage the abilities of different desktop applications in a seamless way It encourages collaboration between applications, each being specialized in a particular task This approach enables the user to assemble a suite of tools according to his personal requirements, combining functionalities of different specialized applications such as VisIVO, Aladin, TOPCAT Through PLASTIC, the applications can share data and link views performing the data exploration with different linked views 8.2 Aladin To find, visualize, and analyze the two-dimensional data like the maps of galaxy clusters simulated by the GADGET-2.0 12 To download this tool, go to http://aladin.u‑strasbg.fr/java/nph‑aladin.pl? frame =downloading 13 This txt file is found at http://palantir3.oats.inaf.it/ITVO/Aladin/ITVO.txt 2008 PASP, 120:933–944 TVO ARCHIVE FOR COSMOLOGICAL SIMULATIONS 943 FIG 12.—Temperature graphics of theoretical galaxy cluster created loading a VOTable downloaded before from the ITVO@Trieste Web portal in the TOPCAT tool See the electronic edition of the PASP for a color version of this figure CONCLUSION This work includes the first prototypes to store, access, cutout, and analyze the cosmological simulation data so as to reuse a very expensive result like big theoretical outputs This work allows scientists to access and compare theoretical and 2008 PASP, 120:933–944 observational data in an easy and homogeneous way using IVOA standards The authors wish to thank Mrs Luigia Santagati for the revision of the English text 944 COSTA ET AL REFERENCES Antonuccio-Delogu, V., Becciani, U., & Ferro, D 2003, Comput Phys Commun., 155 (2), 159 Bonnarel, F., et al 2000, A&AS, 143, 33 Borgani, S., et al 2004, MNRAS, 348, 1078 Comparato, M., et al 2007, PASP, 119, 898 Costa, A., et al 2005, Comput Phys Commun., 166 (1), 17 Costa, A., Becciani, U., Comparato, M., & Gheller, C 2006a, ASP Conf Ser 351, Astronomical Data Analysis Software and Systems XV, ed C.Gabriel, C.Arviset, D.Ponz, & E.Solano (San Francisco: ASP), 445 Costa, A., Becciani, U., Gheller, C., Comparato, M., & Larrson, B 2006b, 26th meeting of the IAU, Special Session 3, The Virtual Observatory in Action: New Science, New Technology, and Next Generation Facilities, 69 Faber, S M., et al 1997, AJ, 114:1771 Gheller, C., et al 2006, Simple Numerical Access Protocol (SNAP) for Theoretical Data (International Virtual Observatory Alliance), http:// www.ivoa.net/internal/IVOA/IVOATheorySNAP/Snap‑0.5‑1.doc Gingold, R A., & Monaghan, J J 1977, MNRAS, 181:375 Hanisch, R J., & Quinn, P J 2003, IVOA Standards (International Virtual Observatory Alliance), http://www.ivoa.net/pub/info/ TheIVOA.pdf, Holmberg, E 1941, ApJ, 94, 385 Lemson, G., et al 2006, preprint (Astro-Ph/0608019) Lemson, G., et al 2007, Data Model for Theoretical (Meta-)data (International Virtual Observatory Alliance), http://www.ivoa.net/ Documents/latest/SimDBTrack.html Lucy, L B 1977, AJ, 82:1013 Manzato, P., et al 2007, in Proc EFMCSC series, Modelling and Simulation in Science, ed V Di Gesù, G Lo Bosco, & M C Maccarone (Singapore: World Scientific), 300 Norman, M L., & Bryan, G L 1999, in Numerical Astrophysics, eds S Miyama, K Tomisaka, & T Hanawa (Heidelberg, Germany: Springer), 19–28 Pasian, F., Taffoni , G., & Vuerli, C 2006, Proc IAU 2, Highlights of Astronomy 14, ed K A van der Hucht (Cambridge: Cambridge Univ Press), 60110.1017/S1743921307011994 Springel, V., et al 2001, NewA, 6, 79 Taffoni, G., et al 2006, preprint (Astro-Ph/0605165v1) Taylor, M B 2005, ASP Conf Ser 347, Astronomical Data Analysis Software and Systems XIV, ed P.Shopbell, M.Britton, & E.Ebert (San Francisco: ASP), 29 Taylor, J., et al.2006, PLASTIC—a Protocol for Desktop Application Interoperability (International Virtual Observatory Alliance), http:// www.ivoa.net/Documents/latest/PlasticDesktopInterop.html 2008 PASP, 120:933–944 ... with the format (HDF5 and GADGET at the moment in our case), resolution, and redshift of the output file The next level includes the link to the code used for the postprocessing, for example, the. .. “Password-Retrieving Form” (asking for the user name) is shown After the submission, an email will be sent to the user to inform him about the password-retrieving procedure The “ITVO Level 0” link brings up a form... contains only the metadata of postprocessing products of the simulation Level of the database includes the description of the FIG 4.—Spherical SNAP form See the electronic edition of the PASP for a color