{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:I5MQMOLERYWF7DW5X2NR47NNEO","short_pith_number":"pith:I5MQMOLE","schema_version":"1.0","canonical_sha256":"47590639648e2c5f8eddbe9b1e7dad23bde21ee4679a62f09eef0327240aa998","source":{"kind":"arxiv","id":"2504.08784","version":1},"attestation_state":"computed","paper":{"title":"SLOs-Serve: Optimized Serving of Multi-SLO LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Arvind Krishnamurthy, Phillip B. Gibbons, Samira Khan, Siyuan Chen, Zhipeng Jia","submitted_at":"2025-04-05T17:41:26Z","abstract_excerpt":"This paper introduces SLOs-Serve, a system designed for serving multi-stage large language model (LLM) requests with application- and stage-specific service level objectives (SLOs). The key idea behind SLOs-Serve is to customize the allocation of tokens to meet these SLO requirements. SLOs-Serve uses a multi-SLO dynamic programming-based algorithm to continuously optimize token allocations under SLO constraints by exploring the full design space of chunked prefill and (optional) speculative decoding. Leveraging this resource planning algorithm, SLOs-Serve effectively supports multi-SLOs and mu"},"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":"2504.08784","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.DC","submitted_at":"2025-04-05T17:41:26Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3d2a4d1e1f3d20a154e5e2a3b135a86c5ee8b48673d5a71090c607da01f5c12a","abstract_canon_sha256":"cb9ee57f1d119ec88bef5d143a41bc2f62a06ec0df7907cd92fd9932d22a4606"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:47:59.344820Z","signature_b64":"xsRyFDuvHPhnYiQAWZNEQhGKZYHeehSS9+Fo4QK+Y5RzIKbJejx51TpbH4Flqb3l4fmRpbkQ/71uvVCHP1/sAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"47590639648e2c5f8eddbe9b1e7dad23bde21ee4679a62f09eef0327240aa998","last_reissued_at":"2026-07-05T10:47:59.344344Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:47:59.344344Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SLOs-Serve: Optimized Serving of Multi-SLO LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Arvind Krishnamurthy, Phillip B. Gibbons, Samira Khan, Siyuan Chen, Zhipeng Jia","submitted_at":"2025-04-05T17:41:26Z","abstract_excerpt":"This paper introduces SLOs-Serve, a system designed for serving multi-stage large language model (LLM) requests with application- and stage-specific service level objectives (SLOs). The key idea behind SLOs-Serve is to customize the allocation of tokens to meet these SLO requirements. SLOs-Serve uses a multi-SLO dynamic programming-based algorithm to continuously optimize token allocations under SLO constraints by exploring the full design space of chunked prefill and (optional) speculative decoding. Leveraging this resource planning algorithm, SLOs-Serve effectively supports multi-SLOs and mu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.08784","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/2504.08784/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":"2504.08784","created_at":"2026-07-05T10:47:59.344404+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.08784v1","created_at":"2026-07-05T10:47:59.344404+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.08784","created_at":"2026-07-05T10:47:59.344404+00:00"},{"alias_kind":"pith_short_12","alias_value":"I5MQMOLERYWF","created_at":"2026-07-05T10:47:59.344404+00:00"},{"alias_kind":"pith_short_16","alias_value":"I5MQMOLERYWF7DW5","created_at":"2026-07-05T10:47:59.344404+00:00"},{"alias_kind":"pith_short_8","alias_value":"I5MQMOLE","created_at":"2026-07-05T10:47:59.344404+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08565","citing_title":"SMetric: Rethink LLM Scheduling for Serving Agents with Balanced Session-centric Scheduling","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07985","citing_title":"Dooly: Configuration-Agnostic, Redundancy-Aware Profiling for LLM Inference Simulation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2508.15919","citing_title":"HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07985","citing_title":"Dooly: Configuration-Agnostic, Redundancy-Aware Profiling for LLM Inference Simulation","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO","json":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO.json","graph_json":"https://pith.science/api/pith-number/I5MQMOLERYWF7DW5X2NR47NNEO/graph.json","events_json":"https://pith.science/api/pith-number/I5MQMOLERYWF7DW5X2NR47NNEO/events.json","paper":"https://pith.science/paper/I5MQMOLE"},"agent_actions":{"view_html":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO","download_json":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO.json","view_paper":"https://pith.science/paper/I5MQMOLE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.08784&json=true","fetch_graph":"https://pith.science/api/pith-number/I5MQMOLERYWF7DW5X2NR47NNEO/graph.json","fetch_events":"https://pith.science/api/pith-number/I5MQMOLERYWF7DW5X2NR47NNEO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO/action/storage_attestation","attest_author":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO/action/author_attestation","sign_citation":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO/action/citation_signature","submit_replication":"https://pith.science/pith/I5MQMOLERYWF7DW5X2NR47NNEO/action/replication_record"}},"created_at":"2026-07-05T10:47:59.344404+00:00","updated_at":"2026-07-05T10:47:59.344404+00:00"}