# Stateful Worlds, Stateless Elasticity: Exact-State Serving for Interactive World Models

arXiv: https://arxiv.org/abs/2607.10389 · HTML: https://mlsanity.com/paper/2607.10389/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 74 references that are in the map.
- First posted on arXiv 11 Jul 2026.

## Builds on (most cited first, 27 of 27)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization](https://mlsanity.com/paper/2401.18079/) (2024)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [LongLive: Real-time Interactive Long Video Generation](https://mlsanity.com/paper/2509.22622/) (2025)
- [CacheGen: KV Cache Compression and Streaming for Fast Large Language Model Serving](https://mlsanity.com/paper/2310.07240/) (2023)
- [Llumnix: Dynamic Scheduling for Large Language Model Serving](https://mlsanity.com/paper/2406.03243/) (2024)
- [BurstGPT: A Real-world Workload Dataset to Optimize LLM Serving Systems](https://mlsanity.com/paper/2401.17644/) (2024)
- [xDiT: an Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism](https://mlsanity.com/paper/2411.01738/) (2024)
- [StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generation](https://mlsanity.com/paper/2511.07399/) (2025)
- [Silent Data Corruptions at Scale](https://mlsanity.com/paper/2102.11245/) (2021)
- [WorldKV: Efficient World Memory with World Retrieval and Compression](https://mlsanity.com/paper/2605.22718/) (2026)
- [Revisiting Network Support for RDMA](https://mlsanity.com/paper/1806.08159/) (2018)
- [X-Cache: Cross-Chunk Block Caching for Few-Step Autoregressive World Models Inference](https://mlsanity.com/paper/2604.20289/) (2026)
- [Execution-State Capsules: Graph-Bound Execution-State Checkpoint and Restore for Low-Latency, Small-Batch, On-Device Physical-AI Serving](https://mlsanity.com/paper/2606.20537/) (2026)
- [CRIUgpu: Transparent Checkpointing of GPU-Accelerated Workloads](https://mlsanity.com/paper/2502.16631/) (2025)
- [GENSERVE: Efficient Co-Serving of Heterogeneous Diffusion Model Workloads](https://mlsanity.com/paper/2604.04335/) (2026)
- [Bit-Exact AI Inference Verification Without Performance Tradeoffs](https://mlsanity.com/paper/2606.00279/) (2026)
- [FlowPrefill: Decoupling Preemption from Prefill Scheduling Granularity to Mitigate Head-of-Line Blocking in LLM Serving](https://mlsanity.com/paper/2602.16603/) (2026)
- [WorldCache: Content-Aware Caching for Accelerated Video World Models](https://mlsanity.com/paper/2603.22286/) (2026)
- [TurboServe: Serving Streaming Video Generation Efficiently and Economically](https://mlsanity.com/paper/2606.19271/) (2026)
- [Demystifying Numerical Instability in LLM Inference: Achieving Reproducible Inference for Mission-Critical Tasks with HEAL](https://mlsanity.com/paper/2606.21023/) (2026)
- [Fast Integrity Verification for High-Speed File Transfers](https://mlsanity.com/paper/1811.01161/) (2018)
- [Adaptive Resource Management and Quality Control for Streaming Video Generation](https://mlsanity.com/paper/2606.15319/) (2026)
- [ShuntServe: Cost-Efficient LLM Serving on Heterogeneous Spot GPU Clusters](https://mlsanity.com/paper/2606.18600/) (2026)

## Abstract (arXiv)

A persistent interactive world model keeps its running state resident on the GPU that serves it: a multi-gigabyte attention cache, almost all of it rewritten at every generation step. That state cannot be recomputed in interactive time or approximated without changing the world, so a live session pins its device. The pin is a scheduling problem. WorldMove moves a live session under one guarantee: the destination is bit-identical to the source, or nothing is installed. It relocates the cache in 18.8 ms same-node, 101x faster than save/load. It holds a checksum-verified 92.1-94.8 Gb/s on a 100 Gb fabric. At that rate the cache fits inside one interactive block. Migrating an actively generating session, it converges at a block boundary and the destination continues the world bit for bit. An admissibility condition decides each move. The move must complete inside the readout horizon, over bandwidth that covers the state plus its dirty rate. Lifted to a fleet schedulability test, it governed a consolidation loop that executed 48 of 48 migrations bit-identical across two providers. Two constraints are structural. Bit-exactness survives only inside a controlled configuration of one GPU architecture, so moving the state is the only way to preserve it exactly in interactive time. Verification cannot hide inside the wire on this fabric. Receive-path checksums stall the transport at protocol timescales under fan-in, and unscheduled incast silently collapses a receiver while every delivered byte stays correct. An incast-aware admission controller holds zero misses to 1.4x offered load and sheds overload as rejects. A lossless GPU codec widens the admission gate to fabrics raw motion cannot use. We exercise the serving loop and the mover separately, each end to end. Their composition on one fabric is unbuilt. Exact-state elasticity is a joint scheduling problem over transport and verification.