What fits local-first cooperation#

The following use-cases benefit greatly from local-first cooperation:

Each of these use-cases lists some useful techniques for that context, here’s the glossary.

What doesn’t#

The use-cases below do not benefit from local-first cooperation. We show them here to give you a better intuition of the scope of this paradigm.

  • online gaming platform: since online games require constant connection and there needs to be a single source of truth for the virtual world, this is best implemented in cloud-native fashion

  • commodities exchange: settling transactions between offers and bids intrinsically requires consensus and is best implemented as a single processing loop with hot stand-by

  • online shop and order fulfilment: the relevant data belong to the store owner, it is a classical client–server setup with clearly distinguished roles, best implemented in cloud-native fashion

  • real-time data analytics for IoT: there is a part that fits greatly, namely showing the current status of devices in the neighbourhood; the larger case of analysing vast volumes from millions of devices and applying machine learning is better implemented in cloud-native fashion, though, using the economies of scale provided there

  • video or photo editing: this is a classical case for standalone programs running on a single computer — only the resulting files are shared with others afterwards

  • low-level machine control: this is best done in hard real-time languages on specialised hardware, no data agency or ownership issues arise at this level (the next level up fits very well, though)

  • crypto currency wallets: all transactions require acknowledgement by the global (and in this sense central!) blockchain, crypto currencies currently don’t work offline

Where the jury is not yet in#

The following cases may benefit from local-first cooperation in principle, but it remains to be proven in practice to present a clear-cut case.

  • video streaming platform: peer-to-peer dissemination of popular content would be extremely efficient, but digital rights handling presents a challenge

  • video conferencing: conceptually it makes a lot of sense to interact directly with one another; this would be easily possible with the early internet’s promise of end-to-end connectivity; conference calls with more than a handful participants need sophisticated video stream dissemination, though, since sending all streams separately to everyone wastes too much upload bandwidth