Manifest are currently used to indexing a content: the transport uses manifests to retrieve the names of the application data and then download them.

It may happen that at a current time point, the transport protocol does not have enough information to fill a window of interests, because of missing manifests. In this case, the transport needs additional information on how to download missing manifests.

The current implementation assumes all the manifests to be equally filled, which is true for a standalone production of a content of a given size. However, it may happen that a producer calls produce several times for different sized contents, resulting in a manifest distribution which is not regular. In this case, there are few options:

• Different namespace for the manifests, with predictive numbering (e.g. sequential)
• Predictive indexing with possibility to receive manifest/content

This patch simplifies the manifest retrieval by modifying the INCREMENTAL strategy in order to download manifest and data with a single incremental sequencing, without waiting manifests for filling the window of interests. Names are known a priori thanks to the NEXT_SEGMENT_CALCULATION_STRATEGY contained in the manifest. This allows to always fill the window and avoiding problems of manifests which are not totally filled.