Update on the IC Roadmap: August & September 2022 Summary

We continue with our tradition of monthly roadmap updates (see June and July edition). This edition covers both August and September, summarizing what we presented during this week’s public global R&D.

In our last update, we announced that we will change our roadmapping format. We discontinued milestones and instead presented stack-ranked lists of features that we work on. We mark features that address community requests and communicate which percentage of our R&D team work in which area. We are convinced that this new approach provides more transparency as to what the foundation works on and when the community can expect what.

In the meantime, we made good progress, even though August is a month where many people take holidays. You can find the new roadmap on internetcomputer.org/roadmap. Please note that the roadmap is and will always be work in progress. Priorities change, things get added or removed. We continue to engage with the community to gather feedback on parts of the roadmap. For example, see @dfx-json 's post ‘22 and ‘23 SDK themes. In parallel, our R&D teams have been preparing bottom-up what they consider important features and tasks to work on next. We continuously update the roadmap accordingly.

Here is an update on recent achievements and upcoming releases.

Core Protocol


  • Internet Identity now supports custom domains. We are still working on a mainstream solution for custom domains but even today, developers can use solutions such as @jplevyak 's IC front to serve their dapp from a custom domain and reuse the same II identities.
  • A few days ago, we switched all boundary nodes to IC OS virtual machines (VMs). The same VMs used for replica nodes. This is a key precondition to eventually handover control of boundary nodes to the NNS…
  • Also beyond the VMs we made good progress with respect to boundary node decentralization. We are about to publish the boundary node target architecture and look forward to your feedback.

System Utilities and Dapps

  • The GIX team updates the NNS FE dapp on a regular basis. These updates include many improvements. One that I appreciate every day is the asynchronous handling of NNS voting, a significant UI improvement.
  • For the last months we participated in a working group to define ICRC-1, a community-supported token standard. A first draft version was re-worked after constructive feedback from the working group. The second draft version was overwhelmingly approved and finally also clearly adopted by the NNS.


  • In this domain, the main focus is the launch of the SNS. @Pete , Daniel T. and @peterparker presented an extensive demo during this week’s public global R&D. It’s exciting to see how all the pieces are starting to come together.
  • One of these building blocks is the community fund. The community fund enables neuron holders to invest in the tokenization of dapps. The community had mixed feelings about the community fund but ultimately the NNS proposal was narrowly adopted and we started working on it.


  • Last month, we presented a set of developer experience (DX) themes that the SDK team is focussing on. There were further discussions on Discord and the forum to collect community feedback on those themes.
  • @jzxchiang asked for more details on Motoko. While we still discuss which larger roadmap items to tackle next, @claudio and @kentosugama presented during this week’s public global R&D the key Motoko extensions and improvements released during the past months.
  • The Developer-facing documentation was opened up so the community can add SDKs, Agents, and other tooling.

Infrastructure & Operations

  • The team just completed tests of the next generation replica node hardware. We focussed on one vendor for now and @garym presented the test results on the forum. We ask the community to help us prioritize which vendors to test next. The node hardware wiki page has also been updated accordingly.
  • The team has made a major effort to switch our build system to Bazel. This leads to build times that are up to 11 times faster compared to today. This is a huge improvement for everybody who builds the IC code regularly.

I hope you share my excitement about all these great achievements. Looking forward to your feedback.


Thank you for your complete explanations


:heart: Thank you for your hard work and great contributions to building a better internet computer. This clear monthly update keeps us on track :smiling_face_with_three_hearts::kissing_heart::rocket:


Thanks! As discussed here, infrastructure and ops roadmap is still highly confusing.

1 Like

@dfisher acknowledged. We hear folks loud and clear on this so we are working on clearing it up.


Will merging harware wallet neurons also be incorporated?

Hi guys!
On the roadmap page in the core protocol section it says:
“Roughly 43 percent of the technical contributions from the R&D team at the DFINITY Foundation revolve around making the Internet Computer protocol efficient, scalable, trustless and interoperable with other blockchains.”
I think there is an error here.

I am embarrassed to admit that I thought “trustless” was intentional, but I do not know its intent (or even really what it means). So possible it is an error. Let me ask @samuelburri

Hi @Azat and @diegop : I think “trustless” makes sense if correctly interpreted. But your question shows that it can also be confusing. My understanding of “trustless” in this context is that you don’t need to trust a single party, i.e. you are not at the mercy of one big player. However, I guess you can also read it as “one cannot trust this protocol”. That was certainly not the intention. I will make a PR to update the page to avoid this confusion.

1 Like

@samuelburri I’m very interested in to learn more about the canister forking feature seen in the “later” row of the Core protocol column. It was also mentioned here by Dominic (starts at 11:10).

I’m specifically interested in the scope of this feature, and some initial thoughts on what changes will need to be made to the protocol in order to support this (high level details/potential engineering approaches).

Having Efficient ways to clone a canister’s data would make it significantly easier to re-partition canisters and their data.

Also, with the upcoming DTS feature, this could allow the user to dynamically add secondary indexes to a live data set (and not have to anticipate all application use cases and data access patterns from the beginning).


Hey @diegop any update here? Appreciate your help in clarifying the plan