CloudView indexing relies on 3 phases: push, analysis and import. In previous versions, the Push Server's role was to receive pushed documents, place them in a task queue, trigger document analysis using the Analyzer, and then send documents to Index Builder for indexing. At best, you could achieve a 10-second latency because the documents were serialized 3 times to disk.

In V6R2014, we've replaced the Push Server, Analyzer and Index Builder processes with the Indexing Server. Instead of using a disk-based task queue, the Indexing Server receives documents, triggers their analysis on the fly, and sends them to the index entirely in memory. This dramatically increases index throughput, and reduces indexing latency to the sub-second level. This new design does not impact the Push-API.

NOTE: If needed, you can still enable a task queue. It's a build group option. When migrating from previous versions, this option will be enabled automatically.

For more details, see "Indexing Process" and "Configuring Indexing" in the CloudView Technical Reference Guide.

Semantic processing often requires resources such as thesauruses, synonyms or blacklists. In V6R2013, we introduced the Resource Manager to manage & deploy these resources on single and multi-host installations, but it was only available on the command line.

In V6R2014, you can now manage resources through the UI:

The Business Console provides a simple XML editor for all resources, as well as rich editors for ontologies, synonyms and rule-based matchers.

Plus, you can test your resources by seeing how sample text gets annotated.

The dictionary keeps track of all the words and their frequency in every indexed document. The dictionary drives spell-checking, regexp matching, related terms as well as corpus-based relevance calculations (Tf/Idf).

In V6R2014, you can now create multiple dictionaries, and configure individual prefix handlers to target a specific dictionary. This reduces the noise that can occur in a single dictionary, which greatly improves quality for spell-check, related terms, and so on. You can further improve quality by defining rules & policies for keeping or discarding words.

For more details, see "Configuring Dictionaries" in the CloudView Technical Reference Guide.

A CloudView Search Based Application includes a frontend and backend configuration, plugins (connectors, widgets...) and resources (thesauruses, CSS, JS, images, ...). 
Now you can use the App packager to bundle all these into a single package, then deploy to another CloudView instances. Variables are available to simplify deploying to different environments (dev, pre-prod, test, production).

For more details, see "Deploying Applications on a CloudView instance" in the CloudView Installation & Deployment Guide.

Jump right to the relevant Administration Console screen thanks to the new search bar. Just start typing the feature name to see the suggested components. It also searches elements you created, like connectors and search logics.

Mashup Builder's now allows you to control the layout within container widgets, such as tabs.

It also introduces new analytic widgets, including scatter-plots, spider and radar charts.

We believe you should be able to select the technology that best suits your needs: in addition to the existing Google Map widget, we now support maps from OpenStreetMap and Bing. Just pick your favorite service!

Identify and fix bottlenecks with the Firebug-like Mashup reporting feature. Once enabled, a timeline provides you with everything you need to profile feed and widget rendering performance. Shave off milliseconds to make your end-users happier.

For more details, see "Enabling Reporting" in the Mashup Builder User Guide.