Just-in-time packaging

The classical OTT workflow encodes once and packages once per delivery format: an MP4 mezzanine becomes a set of CMAF segments, plus HLS playlists, plus DASH manifests, plus separate DRM-encrypted variants. All these artifacts are pre-generated and stored before any viewer arrives. Just-in-time packaging (JIT) inverts this: store only one canonical asset (typically CMAF segments encrypted with one CENC key), and generate HLS playlists, DASH manifests, multi-DRM variants and even ad-insertion stitches at request time.

The benefit is storage: instead of 4× duplication for HLS + DASH × clear + DRM, you store 1×. For a large OTT catalog with hundreds of thousands of titles this can be the difference between 5 PB and 20 PB of object storage. The cost is CPU at the packaging layer — every manifest fetch runs a packaging operation, every segment fetch may apply DRM encryption on the fly. JIT packagers (Unified Streaming, AWS MediaPackage, Akamai's JIT module, EZDRM) are built specifically for this fast-path operation.

JIT is most valuable for catalogs where storage cost dominates CPU cost, and for live-to-VOD workflows where the asset is created in real time but should be available in every format the moment it lands. JIT also enables per-user manifest variants — different bitrate ladders per device, dynamic SSAI stitching per session, personalised renditions — that pre-packaged workflows can't do.

The book · Volume 2 of 7

Video Streaming: A Complete Guide to Video Delivery: From Network Protocols and Transport to Ingest, HLS, and MPEG-DASH

Just-in-time packaging cuts storage by building HLS and DASH outputs on request, at the cost of compute at the origin. Volume 2 covers it in detail: packaging, CMAF, segment protocols such as HLS and MPEG-DASH, and how packaged output reaches players. Written by Nikolay Sapunov, CEO at Fora Soft.

Read it on Amazon →Kindle & paperback · Fora Soft Video Engineering Handbook