Long Term R&D: Scalability (proposal)

Thanks for your questions, @icme !

I’d love to have time to start working on these topics, but right now the teams that would mostly be involved in this are focussing on BTC integration and HTTP calls from canisters.

Note that these Long Term Proposals are really multi-year projects and due to our limited number of people, we are not tackling all of them in parallel.

As you’re probably well aware of, the overall IC throughput of update and query calls can be increased linearly by adding more subnets and node respectively.
There are hard latency and per-subnet update call throughput limits imposed by the bandwidth and geographical distance between the DCs and the query throughput limits depend mostly on the node machine HW. These limits apply to all BFT smart contract platforms.
If you look at Internet Computer performance - Internet Computer Wiki, you’ll see that the IC features very good performance numbers already.
A detailed analysis or projection to how close we can approach the limits has not been performed, as everyone is working on other topics at the moment.

clone canisters in order to scale horizontally
Maybe @johan can chip in with information on what is being discussed in this direction or point to another forum thread?

3 Likes