{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:I2WXYV2SHF5BNI454KGEHMM3RX","short_pith_number":"pith:I2WXYV2S","schema_version":"1.0","canonical_sha256":"46ad7c5752397a16a39de28c43b19b8dd7d8b0d2f222ca36317676a59b2f56da","source":{"kind":"arxiv","id":"2606.02982","version":1},"attestation_state":"computed","paper":{"title":"DriftSched: Adaptive QoS-Aware Scheduling under Runtime Token Drift for Multi-Tenant GPU Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC","cs.LG"],"primary_cat":"cs.PF","authors_text":"Kathiravan Palaniappan","submitted_at":"2026-06-02T00:39:31Z","abstract_excerpt":"The rapid growth of large language model (LLM) inference services has increased the demand for efficient multi-tenant GPU scheduling. While modern inference runtimes such as vLLM improve throughput through continuous batching and optimized memory management, accurately estimating the runtime cost of heterogeneous inference requests remains a significant challenge. In practice, observed output lengths often deviate from admission-time estimates, creating runtime token drift that can lead to workload misclassification, queue imbalance, increased tail latency, and degraded Quality-of-Service (QoS"},"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":"2606.02982","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.PF","submitted_at":"2026-06-02T00:39:31Z","cross_cats_sorted":["cs.DC","cs.LG"],"title_canon_sha256":"0cd6467d9a17255cf04089f92dccc6f49a2bf49b29831a503846fa8d000feb97","abstract_canon_sha256":"0c16e42889603edd59f3842bab285a5657d8d60565101767deae9cde3875924e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:28.449849Z","signature_b64":"In9Fv+VMRRs3g+1jOlVYfFASRy+ZMSlBjUNGex45j2SLx9MwZMtUltZZXKxUaMLh8RE7JPIQGJdjwb4xzHOVBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46ad7c5752397a16a39de28c43b19b8dd7d8b0d2f222ca36317676a59b2f56da","last_reissued_at":"2026-06-03T01:05:28.449422Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:28.449422Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DriftSched: Adaptive QoS-Aware Scheduling under Runtime Token Drift for Multi-Tenant GPU Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC","cs.LG"],"primary_cat":"cs.PF","authors_text":"Kathiravan Palaniappan","submitted_at":"2026-06-02T00:39:31Z","abstract_excerpt":"The rapid growth of large language model (LLM) inference services has increased the demand for efficient multi-tenant GPU scheduling. While modern inference runtimes such as vLLM improve throughput through continuous batching and optimized memory management, accurately estimating the runtime cost of heterogeneous inference requests remains a significant challenge. In practice, observed output lengths often deviate from admission-time estimates, creating runtime token drift that can lead to workload misclassification, queue imbalance, increased tail latency, and degraded Quality-of-Service (QoS"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02982","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/2606.02982/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":"2606.02982","created_at":"2026-06-03T01:05:28.449488+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02982v1","created_at":"2026-06-03T01:05:28.449488+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02982","created_at":"2026-06-03T01:05:28.449488+00:00"},{"alias_kind":"pith_short_12","alias_value":"I2WXYV2SHF5B","created_at":"2026-06-03T01:05:28.449488+00:00"},{"alias_kind":"pith_short_16","alias_value":"I2WXYV2SHF5BNI45","created_at":"2026-06-03T01:05:28.449488+00:00"},{"alias_kind":"pith_short_8","alias_value":"I2WXYV2S","created_at":"2026-06-03T01:05:28.449488+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX","json":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX.json","graph_json":"https://pith.science/api/pith-number/I2WXYV2SHF5BNI454KGEHMM3RX/graph.json","events_json":"https://pith.science/api/pith-number/I2WXYV2SHF5BNI454KGEHMM3RX/events.json","paper":"https://pith.science/paper/I2WXYV2S"},"agent_actions":{"view_html":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX","download_json":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX.json","view_paper":"https://pith.science/paper/I2WXYV2S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02982&json=true","fetch_graph":"https://pith.science/api/pith-number/I2WXYV2SHF5BNI454KGEHMM3RX/graph.json","fetch_events":"https://pith.science/api/pith-number/I2WXYV2SHF5BNI454KGEHMM3RX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX/action/storage_attestation","attest_author":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX/action/author_attestation","sign_citation":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX/action/citation_signature","submit_replication":"https://pith.science/pith/I2WXYV2SHF5BNI454KGEHMM3RX/action/replication_record"}},"created_at":"2026-06-03T01:05:28.449488+00:00","updated_at":"2026-06-03T01:05:28.449488+00:00"}