{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6IN5RTHQ3URNDFN7LHJL6EJAAS","short_pith_number":"pith:6IN5RTHQ","schema_version":"1.0","canonical_sha256":"f21bd8ccf0dd22d195bf59d2bf112004b1044e08c0bb116fbc221f9010979319","source":{"kind":"arxiv","id":"2503.22562","version":1},"attestation_state":"computed","paper":{"title":"Niyama : Breaking the Silos of LLM Inference Serving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Jayashree Mohan, Kanishk Goel, Nipun Kwatra, Ramachandran Ramjee, Ravi Shreyas Anupindi","submitted_at":"2025-03-28T16:04:20Z","abstract_excerpt":"The widespread adoption of Large Language Models (LLMs) has enabled diverse applications with very different latency requirements. Existing LLM serving frameworks rely on siloed infrastructure with coarse-grained workload segregation -- interactive and batch -- leading to inefficient resource utilization and limited support for fine-grained Quality-of-Service (QoS) differentiation. This results in operational inefficiencies, over-provisioning and poor load management during traffic surges.\n  We present Niyama, a novel QoS-driven inference serving system that enables efficient co-scheduling of "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2503.22562","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-03-28T16:04:20Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"d90b060c0cc9b40ce39c73a2744d97926a608052784e0836939b9e9f2db82f3f","abstract_canon_sha256":"cb46404fec1f30c1b306fd940742bc31971e1643556ae761e904a88d2b6c46df"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:41:06.827112Z","signature_b64":"N8ZGiTcSVTB8HkSplwdM3HaTkDoF6f8ukI/teWSJfCxDRDdiRvRL/UvYkPXr0dgEf+JaqeBoZxpAFs3zbR+EDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f21bd8ccf0dd22d195bf59d2bf112004b1044e08c0bb116fbc221f9010979319","last_reissued_at":"2026-07-05T10:41:06.826568Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:41:06.826568Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Niyama : Breaking the Silos of LLM Inference Serving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Jayashree Mohan, Kanishk Goel, Nipun Kwatra, Ramachandran Ramjee, Ravi Shreyas Anupindi","submitted_at":"2025-03-28T16:04:20Z","abstract_excerpt":"The widespread adoption of Large Language Models (LLMs) has enabled diverse applications with very different latency requirements. Existing LLM serving frameworks rely on siloed infrastructure with coarse-grained workload segregation -- interactive and batch -- leading to inefficient resource utilization and limited support for fine-grained Quality-of-Service (QoS) differentiation. This results in operational inefficiencies, over-provisioning and poor load management during traffic surges.\n  We present Niyama, a novel QoS-driven inference serving system that enables efficient co-scheduling of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.22562","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2503.22562/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2503.22562","created_at":"2026-07-05T10:41:06.826628+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.22562v1","created_at":"2026-07-05T10:41:06.826628+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.22562","created_at":"2026-07-05T10:41:06.826628+00:00"},{"alias_kind":"pith_short_12","alias_value":"6IN5RTHQ3URN","created_at":"2026-07-05T10:41:06.826628+00:00"},{"alias_kind":"pith_short_16","alias_value":"6IN5RTHQ3URNDFN7","created_at":"2026-07-05T10:41:06.826628+00:00"},{"alias_kind":"pith_short_8","alias_value":"6IN5RTHQ","created_at":"2026-07-05T10:41:06.826628+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18431","citing_title":"Beyond Prediction: Tail-Aware Scheduling for LLM Inference","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07609","citing_title":"Blink: CPU-Free LLM Inference by Delegating the Serving Stack to GPU and SmartNIC","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17227","citing_title":"Cloud-native and Distributed Systems for Efficient and Scalable Large Language Models -- A Research Agenda","ref_index":74,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS","json":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS.json","graph_json":"https://pith.science/api/pith-number/6IN5RTHQ3URNDFN7LHJL6EJAAS/graph.json","events_json":"https://pith.science/api/pith-number/6IN5RTHQ3URNDFN7LHJL6EJAAS/events.json","paper":"https://pith.science/paper/6IN5RTHQ"},"agent_actions":{"view_html":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS","download_json":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS.json","view_paper":"https://pith.science/paper/6IN5RTHQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.22562&json=true","fetch_graph":"https://pith.science/api/pith-number/6IN5RTHQ3URNDFN7LHJL6EJAAS/graph.json","fetch_events":"https://pith.science/api/pith-number/6IN5RTHQ3URNDFN7LHJL6EJAAS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS/action/storage_attestation","attest_author":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS/action/author_attestation","sign_citation":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS/action/citation_signature","submit_replication":"https://pith.science/pith/6IN5RTHQ3URNDFN7LHJL6EJAAS/action/replication_record"}},"created_at":"2026-07-05T10:41:06.826628+00:00","updated_at":"2026-07-05T10:41:06.826628+00:00"}