PLATFORM TO EASE THE DEPLOYMENT AND IMPROVE THE AVAILABILITY OF TRENCADIS INFRASTRUCTURE IberGrid 2013 Miguel Caballer GRyCAP – I3M - UPV

INDEX TENCADIS Technology Motivation Objectives Analysis of TRENCADIS Services Platform Architecture TRENCADIS Cloud infrastructure

TRENCADIS Technology TRENCADIS: Towards a Grid Environment for Processing and Sharing DICOM Objects Aim to share DICOM objects (medial imaging) among different medical centres, including annotation data from DICOM-SR (Structured Reports). Use Standard Components that can be integrated in existing Grid infrastructures (such as EGI or Ibergrid). Provide developers with an object-oriented components distributed in packages to develop final applications and user interfaces for browsing and managing repository content.

TRENCADIS Technology A test infrastructure of TRENCADIS was deployed in 2008 (CVIMO project – 4 public hospitals and a private hospital was involved). Currently, TRENCADIS is deployed in a real infrastructure among two centres  The UPV- I3M-GRyCAP and Dr. Peset Hospital. These Deployments has reported the value as a tool for improving the access to DICOM objects.

MOTIVATION Also, in these deployments were detected that the deployment and maintenance of TRENCADIS services require a considerable investment of time and effort. Mainly because of: The intrinsic complexity of Grid. Standards and systems used in the hospitals. The network security restrictions imposed on the hospitals. This fact makes difficult adapt a deployment in some scenarios, such as an unexpected increase in demand or in the number of images stored in the system. This work was motivated by these problems, providing the basis for new TRENCADIS deployment strategies, using Cloud technologies.

OBJECTIVES Design a platform for the deployment of TRENCADIS- based applications, using Virtualization and Cloud computing techniques. The platform ….. avoids intrusive deployment of services to reduce the amount of effort required to install and maintain new TRENCADIS sites. provides mechanisms to monitor and handle performance and reliability requirements together, allocating and de- allocating different computational resources from the Cloud, as required by the applications. will be designed to be portable to other application domains that also rely on Grid computing.

ANALYSIS OF TRENCADIS SERVICES This infrastructure is composed by a set of services based on Grid technologies which are integrated in a Virtual Organization (VO). There are two categories of Services: CORE services. Located in each hospital involved. SERVER services. Located in a external TRENCADIS center. Scientific Linux 5.7 OS is recommended for all of them, although another compatible GNU/Linux OS could be used

ANALYSIS OF TRENCADIS SERVICES (example: DICOM Storage) An Analysis of TRENCADIS Services has allowed to detect the VMI and software components of each VMI detected. VMIs: Indexer. Index the information. Backend. Kept images and reports. Storage DICOM Grid Service. Grid Service Interface. Software: Indexer. PostgreSQL or AMGA or Neo4J. BackEnd. PostgreSQL or File System or GridFTP or LFC+SE or CDMI. Storage DICOM Grid Service. Java API Indexer, Java API Backend, Globus 4, Java and Ant.

PLATFORM ARCHITECTURE

PLATFORM ARCHITECTURE - RVMI Repository of VMIs (RVMI) Implemented with the VMRC Index and store VMIs together with metadata descriptions OS, applications, etc. Enables sharing and reusing VMIs Repository of VMIs (RVMI) Implemented with the VMRC Index and store VMIs together with metadata descriptions OS, applications, etc. Enables sharing and reusing VMIs

PLATFORM ARCHITECTURE - RSC Repository of Software Components (RSCs) Installations files needed to deploy and configure the infrastructure Repository of Software Components (RSCs) Installations files needed to deploy and configure the infrastructure

PLATFORM ARCHITECTURE - RSD Repository of Service Descriptions (RSDs) Describe the composition and configuration of the services Hardware & Software Using RADL Repository of Service Descriptions (RSDs) Describe the composition and configuration of the services Hardware & Software Using RADL

PLATFORM ARCHITECTURE - II Infrastructure Instantiator (II) REST API & Web Interface Performs the management of the infrastructure using the RADL description Support Cloud and Virtualization systems: EC2, OpenNebula, OpenStack, LibVirt Contacts the RVMIs to select the most suitable VMI Performs a contextualization process using Puppet Infrastructure Instantiator (II) REST API & Web Interface Performs the management of the infrastructure using the RADL description Support Cloud and Virtualization systems: EC2, OpenNebula, OpenStack, LibVirt Contacts the RVMIs to select the most suitable VMI Performs a contextualization process using Puppet

RADL EXAMPLE network private network public (outbound = 'yes') system GT4_JAVA_ANT ( cpu.arch='x86_64' and cpu.count>=1 and memory.size>=1024m and net_interfaces.count = 2 and net_interface.0.connection = 'public' and net_interface.1.connection = 'private' and disk.0.os.name='linux' and disk.0.os.flavour='Scientific Linux' and disk.0.os.version='5.7' and disk.0.application contains (name='globus', version='4') and disk.0.application contains (name='java', version=' ') and disk.0.application contains (name='ant', version='1.8.2') ) system BACKEND_POSTGRESQL ( cpu.arch='x86_64' and cpu.count>=1 and memory.size>=1024m and net_interfaces.count = 1 and net_interface.0.connection='private' and disk.0.os.name='linux' and disk.0.os.flavour='Scientific Linux' and disk.0.os.version='5.7' and disk.0.application contains (name='PostgreSQL', version>='8.4.9') ) configure GT4-JAVA-ANT ( add_user {'trencadis': } deploy_gs {'Key_Server_Grid_Service.gar': } install_conf_API_SQL_Keys_Database {'TRENCADIS_Java_API_SQL_Keys_DB.jar': } ) configure BACKEND_POSTGRESQL( add_user {'trencadis': } create_database {'SQL_Keys': file =>'database.SQL'} ) deploy GT4_JAVA_ANT 1 deploy POSTGRESQL 1

TRENCADIS CLOUD INF.

CONCLUSIONS Designed a platform for the deployment TRENCADIS infrastructures Simplifies the deployment Improving its performance and reliability New services can be provisioned on demand, in an elastic and dynamic way. This methodology can be applied to other infrastructures Identify and apply the methodology in different use cases.