Seekable OCI: Lazy-Loading Container Images via Range-Request Indexing

2026-07-09T08:52:21Z2a35de4fb95b4f828d536b51f5a721ae70c4da8a855be51e98317289592a9305
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.