Seekable OCI: Lazy-Loading Container Images via Range-Request Indexing
2026-07-09T08:52:21Z•2a35de4fb95b4f828d536b51f5a721ae70c4da8a855be51e98317289592a9305
ECO/CPO-DAGECS FargateEKSFP8FUSE filesystemGIFTHTTP range requestsKubernetesNVFP4OCI referrerSOCISeekable OCIVoltroncontainer imagescontradiction proofedge inferenceimage pull performancelazy-loadinglow-precision gradient communicationmodel partitioningmulti-device LLMprivacyselective disclosuresupply chain accountabilityzero-knowledge
What happened
This collection of recent systems and ML papers includes a high-impact container-image optimization (SOCI: Seekable OCI) that enables lazy-loading of standard OCI images via an external index stored as an OCI referrer artifact and served with a FUSE filesystem using HTTP range requests, achieving large cold-start speedups in Kubernetes/ECS. Related work covers edge/elastic LLM inference (Voltron), geometry-aware low-precision gradient communication for LLM pretraining (GIFT), a contradiction-detection accountability layer for adversarial supply chains (ECO/CPO-DAG), DNN compilation optimizaion
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arxiv_cs_dc
- Record identifier
- 2a35de4fb95b4f828d536b51f5a721ae70c4da8a855be51e98317289592a9305
- Enrichment time
- 2026-07-09T08:52:21Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.