↓Skip to main content
alchemiscale.org

alchemiscale roadmap 2025

·645 words·4 mins·
Author
David L. Dotson
Lead Developer // Datryllic LLC
Table of Contents

It’s 2025, and this month the project now known as alchemiscale celebrates its 3rd anniversary since kicking off in 2022. We’ve come a long way since then, and there is much to do this year to further improve alchemiscale’s capabilities and user experience.

We previously published a detailed perspective on what we considered to be the alchemiscale project’s most attractive advancement opportunities in 2025. That document serves as an input to this roadmap, which details our plans for executing on many of those opportunities over the course of the year.

our objectives
#

This year, our overall objectives are:

  1. Automated Strategy pursuit server-side via the Strategist service.

  2. Mechanisms for finer and more performant introspection on Transformation results and failures, and performant whole-AlchemicalNetwork result retrieval.

  3. Improved compute efficiency through better saturation of compute resources, and mechanisms for autoscaling compute services on those resources based on user demand.

  4. Flexible “living networks” support, including merging and copying AlchemicalNetworks server-side, and prediction of AlchemicalNetwork computational cost.

By the end of 2025, accomplishing much of the above would make alchemiscale largely feature-complete as originally envisioned. This would position us well to release alchemiscale v1.0.0, marking a high degree of API stability and greater suitability for production environments.

our resources
#

We perform alchemiscale development via two-week sprints within the OpenFE organization, with development currently led by Datryllic. The OpenFE industry-funded project currently does not allocate its developer resources directly to alchemiscale, so resourcing comes through the following modes:

  1. Service contracts with Datryllic from entities that depend on alchemiscale, either a bespoke engagement for building specific features, or a plug-and-play support engagement that funds continuous overall improvement of alchemiscale.
  2. Contributions of internal developer effort from organizations that are invested in alchemiscale’s future for their own operations.
  3. Individual developer effort, volunteered for personal or professional reasons.

As of this writing, Datryllic remains the primary contributor to alchemiscale, with its work supported via mode (1). We welcome contributions via the other modes in pursuing our objectives over the course of the year, however! Please reach out via our Discussions forum if you are interested in contributing with us!

our path
#

Given our current resources, a high degree of team focus will be required to deliver features aligned with our objectives. Major releases will be structured according to a theme, allowing related features to be grouped together and co-developed in the same release cycle. This also helps us to realize more tangible, coherent improvements with users more frequently.

As of this writing, our next major releases will focus on the following:

  • v0.6.0 : result retrieval optimizations, server-side Task restart policies
  • v0.7.0 : automated Strategy execution via Strategist service
  • v0.8.0 : living networks enablement, fast traceback retrieval, archival extracts
  • v0.9.0 : improved compute efficiency for parallelizable ProtocolDAGs

Our aim is to produce one major release approximately each quarter, prioritizing features that align with each release’s overall theme in service to our objectives. We may rebalance individual features across major releases as needed, but this is the overall principle we use to guide what is triaged into release milestones.

In service to objective (3) Datryllic will also allocate development effort to alchemiscale-k8s, in particular designing and implementing an approach to efficient compute service autoscaling on Kubernetes clusters. The business logic of this approach may end up being reusable for developing an autoscaling approach for HPC and other platforms, but this is yet to be seen.

Datryllic is also committed to advancing use of Folding@Home with alchemiscale via alchemiscale-fah, unlocking citizen-science, “planetary-scale” compute for projects that can make use of it.

onward
#

There’s quite a lot to do this year, and plenty of opportunities for new folks to get involved. If you are interested in trying alchemiscale out, or if you already have and want to offer ideas for improvement, please reach out! Posting in our Discussions forum is the best way to get started!