{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:M6S5LMKC7OR6JGADFHLCWLGXCS","short_pith_number":"pith:M6S5LMKC","schema_version":"1.0","canonical_sha256":"67a5d5b142fba3e4980329d62b2cd71492d7f248f7791a65501d186fd482cb2f","source":{"kind":"arxiv","id":"2505.23022","version":1},"attestation_state":"computed","paper":{"title":"SCORPIO: Serving the Right Requests at the Right Time for Heterogeneous SLOs in LLM Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hui Lu, Tingfeng Lan, Wei Chen, Xiuqi Huang, Yinghao Tang","submitted_at":"2025-05-29T03:03:09Z","abstract_excerpt":"Existing Large Language Model (LLM) serving systems prioritize maximum throughput. They often neglect Service Level Objectives (SLOs) such as Time to First Token (TTFT) and Time Per Output Token (TPOT), which leads to suboptimal SLO attainment. This paper introduces SCORPIO, an SLO-oriented LLM serving system designed to maximize system goodput and SLO attainment for workloads with heterogeneous SLOs. Our core insight is to exploit SLO heterogeneity for adaptive scheduling across admission control, queue management, and batch selection. SCORPIO features a TTFT Guard, which employs least-deadli"},"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":"2505.23022","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-29T03:03:09Z","cross_cats_sorted":[],"title_canon_sha256":"3e2e8987b41228abaa0fdfbf22a36abacb81bfa885724bca49c0cb7f1d2bd0e0","abstract_canon_sha256":"e49414be45ab19ab6003d48d8e6399bc01a0f2d02661ddaa4410bcba5bdd3fd2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:53.717457Z","signature_b64":"Uj4po0tEKSCg71WOR+4P8hGQcHzK6PGCXAhy2gnKXHC1z34pl2zh1fgKiiMVTXFzlusam3TQU/lZ2Sfw06XVBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"67a5d5b142fba3e4980329d62b2cd71492d7f248f7791a65501d186fd482cb2f","last_reissued_at":"2026-07-05T11:11:53.716932Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:53.716932Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SCORPIO: Serving the Right Requests at the Right Time for Heterogeneous SLOs in LLM Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hui Lu, Tingfeng Lan, Wei Chen, Xiuqi Huang, Yinghao Tang","submitted_at":"2025-05-29T03:03:09Z","abstract_excerpt":"Existing Large Language Model (LLM) serving systems prioritize maximum throughput. They often neglect Service Level Objectives (SLOs) such as Time to First Token (TTFT) and Time Per Output Token (TPOT), which leads to suboptimal SLO attainment. This paper introduces SCORPIO, an SLO-oriented LLM serving system designed to maximize system goodput and SLO attainment for workloads with heterogeneous SLOs. Our core insight is to exploit SLO heterogeneity for adaptive scheduling across admission control, queue management, and batch selection. SCORPIO features a TTFT Guard, which employs least-deadli"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.23022","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/2505.23022/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":"2505.23022","created_at":"2026-07-05T11:11:53.716995+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.23022v1","created_at":"2026-07-05T11:11:53.716995+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.23022","created_at":"2026-07-05T11:11:53.716995+00:00"},{"alias_kind":"pith_short_12","alias_value":"M6S5LMKC7OR6","created_at":"2026-07-05T11:11:53.716995+00:00"},{"alias_kind":"pith_short_16","alias_value":"M6S5LMKC7OR6JGAD","created_at":"2026-07-05T11:11:53.716995+00:00"},{"alias_kind":"pith_short_8","alias_value":"M6S5LMKC","created_at":"2026-07-05T11:11:53.716995+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09174","citing_title":"Demonstrating chart-plot: Closing the Last Mile of Academic Chart Generation","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08267","citing_title":"Execution Envelopes: A Shared Admission Contract for Backend AI Execution Requests","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS","json":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS.json","graph_json":"https://pith.science/api/pith-number/M6S5LMKC7OR6JGADFHLCWLGXCS/graph.json","events_json":"https://pith.science/api/pith-number/M6S5LMKC7OR6JGADFHLCWLGXCS/events.json","paper":"https://pith.science/paper/M6S5LMKC"},"agent_actions":{"view_html":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS","download_json":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS.json","view_paper":"https://pith.science/paper/M6S5LMKC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.23022&json=true","fetch_graph":"https://pith.science/api/pith-number/M6S5LMKC7OR6JGADFHLCWLGXCS/graph.json","fetch_events":"https://pith.science/api/pith-number/M6S5LMKC7OR6JGADFHLCWLGXCS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS/action/storage_attestation","attest_author":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS/action/author_attestation","sign_citation":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS/action/citation_signature","submit_replication":"https://pith.science/pith/M6S5LMKC7OR6JGADFHLCWLGXCS/action/replication_record"}},"created_at":"2026-07-05T11:11:53.716995+00:00","updated_at":"2026-07-05T11:11:53.716995+00:00"}