mlsanityData to 14 Oct 2026

The DMA Streaming Framework: Kernel-Level Buffer Orchestration for High-Performance AI Data Paths

arXiv 2603.10030 ↗First posted 26 Feb 2026cs.AR cs.AI cs.DCcited by 0

Not on the map. Why?

Abstract (arXiv)

AI transport libraries move bytes efficiently, but they commonly assume that buffers are already correctly allocated, placed, shared, registered, and safe under completion and teardown pressure. This paper presents dmaplane, a Linux kernel module that makes this missing layer explicit as buffer orchestration. dmaplane exposes a stable kernel UAPI via /dev/dmaplane and composes ring-based command channels, DMA buffer lifecycle management, dma-buf export for cross-device sharing, a kernel-space RDMA engine, NUMA-aware allocation and verification, credit-based flow control, low-overhead observability, and GPU memory integration via PCIe BAR pinning. We evaluate orchestration sensitivity with measurements of NUMA cross-node penalties at DRAM scale, completion-safe flow control under sustained RDMA load, and GPU BAR mapping tiers versus cudaMemcpy. We also demonstrate end-to-end disaggregated inference by transferring KV-cache chunks between two machines using RDMA WRITE WITH IMMEDIATE and reconstructing tensor views on the receiver. RDMA measurements use Soft-RoCE; we distinguish measured results from provider-independent properties by construction.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 5 of its 48 references that are in the map. First posted on arXiv 26 Feb 2026.

Builds on

5 of 48 references.

References · 5

  1. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  2. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,141
  3. 2024DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Servingcited by 617
  4. 2023Splitwise: Efficient generative LLM inference using phase splittingcited by 558
  5. 2025fabric-lib: RDMA Point-to-Point Communication for LLM Systemscited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Understanding Bottlenecks for Efficiently Serving LLM Inference With KV Offloading
  2. 2026The Serialized Bridge: Understanding and Recovering LLM Serving Performance under Blackwell GPU Confidential Computing
  3. 2025DataMaestro: A Versatile and Efficient Data Streaming Engine Bringing Decoupled Memory Access To Dataflow Accelerators
  4. 2026TENT: A Declarative Slice Spraying Engine for Performant and Resilient Data Movement in Disaggregated LLM Serving
  5. 2026DAK: Direct-Access-Enabled GPU Memory Offloading with Optimal Efficiency for LLM Inference ↗
  6. 2026From Tensor Buffer to Distributed Memory Hierarchy: A Survey of KV Cache Management for LLM Serving
  7. 2026KVDrive: A Holistic Multi-Tier KV Cache Management System for Long-Context LLM Inference
  8. 2026CacheFlow: Efficient LLM Serving with 3D-Parallel KV Cache Restoration