Put it against a real system
Architecture and distributed-platform engagements expose the real seams: state, messaging, partitions, recovery, and the operating handoff.
See the practice← Research agenda
01 · Communication
Fleet consistency
Research question
Time to bring a growing, distributed agent fleet back into consistency.
Why the problem is difficult
The mechanism and its failure modes.
The cost of consistency grows faster than the fleet. More agents create more state, more relationships, and more ways for the current picture to split.
Embodiment makes the problem harder again. Hardware adds physical state, latency, intermittent connectivity, and work that cannot pause while a region catches up.
Current state
What Anant is doing now.
The research question and time-to-consistency measure are defined. Tekt provides an open reference surface for edge-to-cloud work, but Anant has not yet published a fleet benchmark, protocol specification, or conformance result. Those are the work, not claims we can make early.
Proposed method, output, and evaluation
Proposed method: model a heterogeneous fleet through new information, partitions, and node rejoining, then measure time to consistent state under each condition. Proposed output: a documented protocol, reference implementation, and evaluation harness. None is being presented as complete today. Evaluation: compare time to consistency, recovery behavior, and message cost across the defined failure conditions.
From question to output
Two ways this work becomes useful.
Produce an inspectable result
The result should apply beyond Anant’s commercial scope. Fund the measurement, specification, implementation, and public artifact—not a consulting subsidy.
Discuss funding