Learn more about how our technology is already being put to use. And how we focus on understanding your business objectives and create according solutions, spanning from local servers to cloud services, from ontologies to billions of free-form items, across continents, connecting with existing infrastructure and getting things done, from scribbles to launch in weeks instead of years. Below you find a selection of existing use-case summaries:

System One Radar indexes up to 1 million news articles from various sources every day. Every article is being analyzed, tagged, clustered and stored in a distributed back-end index currently holding more than 500m documents. Highlights of this implementation are: A highly scalable algorithmic web service based on Sun servers that can be extended on-demand with a large number of hot-deployed Amazon EC2 nodes, eg in case a large amount of data is being reprocessed with improved algorithms. Visualizations render large parts of the dataset in sub-second timeframes and front-end options span from desktop to mobile to paper.

System One Collaboration combines a semantic, wiki-style collaboration platform with federated search. Our SaaS offering spans dedicated servers for the core services, a CDN for static application components as well as encrypted Amazon S3 storage to offer customers unlimited space for files & images. In-house deployments are consolidated on a single server, or multiple for larger instances and/or redundancy. The main algorithmic focus is on content and user profile / interest matching, while the triple blackboard is directly accessible to the user for the creation of 'structured tags' as well as 'smart queries'.

Federated search, named entity recognition and clustering / similarity are the key issues when it comes to internal applications. Most common is a combination of them, where the System One platform acts as unification point of various sources, from file-shares to DMS/KMS systems, offering a single facetted search interface, related information and tags next to every resource as well as the automated creation of topic specific dossier pages, portals and portlets. A more specific example is pictured below: An Expert search engine unifying more than two dozen heterogeneous internal, proprietary and public data sources with a total of more than 600 m documents, normalizing all of them and deriving more than 800 k person profiles from that. Highlights include a rapid prototyping and feedback phase of 2 months and a final implementation within another 6 weeks, on the technical side the project featured a one-time password authentication for public terminal / mobile access based on the truly credit card sized InCards and Vasco back-end as well as single WAR file or VMWare image deployments and a custom analytics and management dashboard.

Most projects in this area start by creating lightweight web service wrappers around often in-house made legacy content management systems. Once this is achieved, a first dump is transferred to the platform and a process with the editorial team is being initiated, in which various data strategies are being jointly defined, tested and refined to reflect the characteristics of the medium. At the same time, workflows are implemented via dedicated management GUIs or plugins / widgets within the existing publishing system. Once this is accomplished final deployments are rolled out with real-time capabilities (meaning little to no caching to stay up to date), all existing archives processed and additional nodes added according to traffic requirements, which can scale well into billions of requests per month.

In this case, traditional Database-CMS-Front approaches are being replaced by a Mashup of various backend web services. Based on existing web services, custom harvesters and wrappers as well as a very thin front-end layer, it's easily possible to create news aggregators, custom topic specific portals or central service clouds unifying scattered existing infrastructure. A typical workflow could mash up a conditioned EPG (TV program) stream and news / social media contents that are semantically enriched (eg tags extracted etc.). Once all the information is unified in the platform, a custom data strategy with underlying algorithms like item clustering and recommendations create a new kind of TV program companion thats available on the web, and via simple skins or API tapping apps also as Desktop, TV overlay or Web widget, mobile interface and IPTV web service infrastructure.