This project explores the possibilities of a ‘second screen’ application. While watching television (first screen), more and more people are simultaneous connected and interacting with the internet using their tablet/computer (second screen).

The MediaSquare application captures the social interaction between the viewer and the show. Think a twitter/facebook like interaction for a television show. Additionally, applications/content relevant to the show are integrated.

In the MediaCRM application, all this interaction is captured in a data-warehouse to act as a customer relationship (CRM) for further analysis. This allows suggestions for viewers based on past behavior, their friends behavior, resulting in targeted information.

An additional dimension was added by encoding an additional (hidden) signal into the broadcast signal, so external devices (Ipad) could auto-detect through the microphone, which TV-show the user is currently watching and the exact position/time in the show. Now the level of interaction could also depend on the content and time in the show: f.i. a quiz could be displayed at only a certain point.

The main difference between this and a traditional website, is that this kind of data collection is extremely bursty: the ramp-up period is much higher and this requires measures for both scaling and performance. The target for performance is a million updates/ 5 seconds. And for the scaling both up and down scaling are important to allow only allocation of resources necessary, making a perfect fit for cloud use.

Virtual machines are more commonly used to segment applications and services. This typically induces ‘Image Sprawl’. The project was to manage the wild grow of these diverging platforms.

• To manage this problem puppet was introduced as a configuration management tool.
• As a side effect, backup licenses were not required anymore for the virtual machines that could be recreated from scratch and did not contain production data. * This totally adhered to the ‘infrastructure as code’ principles.
• In addition to the OS and middleware automation, the creation of the virtual machines and routing on the Vmware ESX was automated using the java vijava API.
• Further integration with the application development cycle was done by integrating maven, nexus and Puppet J2EE war deployment.
• Integration with BigIP loadbalancing during creation of new machines

As the number of machines in the synergics datacenter have grown over the years, the current datacenter was becoming inadequate in terms of cooling and electricity. Therefore part of the infrastructure needed to be migrated to a new location. This involved:

• Selecting and preparing a new housing facility and electricy needs
• Preparing internet and inter-datacenter network connectivity redundancy including firewall and loadbalancer functionality
• Testing, architecting a new storage solution (Nexenta and Netapp) and data migration scenario
• Preparation of new blade HP infrastructure on the new site
• Vmware esx cluster setup and virtual machine migration to the new site
• Migration of application in high availability mode to become loadbalanced across both sites
• Integration in Nagios monitoring

At the same time another project was preparing an upgrade of the Oracle Database server

• upgrade to a new security model, role based access
• upgrade from Oracle 10 to Oracle 11
• setup of test environment for application migration
  

In the past two year the VRT has been feeding a new media archive system with all the new content it is producing. The purpose of this project is to make this data more “searchable”. Combining this new data with the old archive data, new ways of exploring are tested. F.i. faceted searches, timeline, coverflow and so on.

Aside from the content that is handcoded, new ways of extracting metadata like face detection, audio detection are explored to feed even more information into the system

• Exploring new User interfaces for faster retrieval of video items
• Wikipedia as collaborative thesaurus effort instead of VRT thesaurus
• Creation of a Media Asset Management System (MAM) for storing video content
• Large File Upload, Transfer and Storage management
• Collaborative Moodboard,/ideaboard for micro-commissioning new movie Projects(IFIP project)
• Integrate drupal with rails application using REST API
• Publish corporate video and metadata to the public and mining of descriptions (Memento Project)
  

• Version 7.x of the Sun ONE Portal includes several collaboration tools like wiki’s, forums and chat. The purpose of this proof of concept was the integration of these new tools within the existing Eurocontrol environment.

• The integration between a web based portal environment and a thin client environment based upon Sun Secure Global Desktop (Tarantella) was established using portlet and netlet technology.
  

• In order to improve access to the medical data and applications, AZ Nikolaas started a Request for Proposal for a new Identity management System. They had stated their functional requirements in this document.

• Different vendors proposed different solutions with different architectures. The evaluation of the impact and match of these solutions was difficult because they had no experience with the proposed solutions.
  

• Fedict in the past only did contract management. Because of the inflexibility of creating a common platform, the decision was made to migrate all existing applications to a distributed datacenter with dedicated infrastructure. This datacenter will also host the portal of Belgium and all inter government SOA communication hubs.
• In order to migrate several internal processes need to be implemented, both in service delivery as in service management. ITIL was used as the preferred methodology. Still the methodology is only a guidance, the whole company was changed in order to handle this.
• Due to the complex nature of the applications and lack of clear and current documentation, also the project methodology slowly embraced a more iterative approach: timeboxing, iterative design and dedicated team were chosen to better deliver in time of the ‘political’ milestones.
• The datacenter technology was based upon Redhat Xen virtualization, resulting in separated but centrally managed infrastructure. Installation and configuration of not only the operating system but also the Oracle and Bea Weblogic was automated and integrated in the test cycle.
  

• Telenet as a telecom operator , has a lot of partners integrating with internal processes.
  The first signs of user management were solved using custom inhouse built solutions and have integration with ADAM/Active Directory. The scope of this project is to replace these systems with an industry based identity and access management and open solution.

• The OpenSSO solution is based upon Sun Access Manager solution. The internal application landscape demanded a more distributed policy management.

• Credoc technically runs the portal and mail environment used by every notary in Belgium. (+/- 8000). As Trust is very important they have high security standards to implement. After an Audit which they successfully past, the next step was to improve the reliability of the solution even more by implementing an Disaster Recovery Solution.
• While Credoc technically passed the Audit, the most urgent was not a technical one but an organizational: Operations successfully ran the environment but too many new projects pushed things into production. Given the understaffing, this resulted in more operational errors and incidents increased and less time was available for the Disaster Recovery Plan.

• Using Scrum both the operational Systems group was aligned with the operational Application group. On a higher level the projects became aware that the operational group was a ‘shared’ resources and therefore a bottleneck in the organization. Priorities were put in a Program Plan. A new ticketing system was introduced to better overview ongoing efforts.

• On the technical level , the architecture was designed based upon the multisite concept and blades systems of HP.
• For better and faster deployment of projects, different environments were set up and the whole process was documented and further automated, using continuous integration.
  

• The railway company “NMBS” made a Request For Proposal for Identity and Access Management.
• One of the major challenges was the introduction of disaster recovery capabilities in the Identity and Access Management solution. Due to the distributed nature of their enduser base this resulted in high failover clustering systems.
• The proposal also provided a solution for creating different environments for development, testing, integration, production and disaster recovery. Hardware was virtualized using Solaris zones technology.
  

• Initial scope was just to put the redundant hardware in a second location. While this would solve solve problems, there was a lot coordination to be done within each layer of the identity and access management solution. Each subcomponent provided its own redundancy layer but no orchestration was foreseen between the different layers.
• An example was that f.i. The loadbalancers was still active on one site, the ldap on the second, the database again on the first. Different coordination strategies were described and integrated in the solution.
• Another aspect was log consolidation and data loss and integrity within disaster recovery solution. The right balance between consolidation and distribution was analyzed. Due to the financial impact of the information additional encryption was foreseen at network Layer 2 and Layer 3. Also at the storage (SAN) level additional measures were taken.
• The different disaster recovery strategies for the applications using the access control mechanisms were examined and the impact on the architecture was studied.