{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:Y4EOI4VS4I4B7DDUME5IGWMWH2","short_pith_number":"pith:Y4EOI4VS","schema_version":"1.0","canonical_sha256":"c708e472b2e2381f8c74613a8359963e88561ca32dabbebca9f4fa6e47055d49","source":{"kind":"arxiv","id":"2403.01876","version":1},"attestation_state":"computed","paper":{"title":"D\\'ej\\`aVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Amar Phanishayee, Ana Klimovic, Foteini Strati, Jakub Tarnawski, Sara McAllister","submitted_at":"2024-03-04T09:32:05Z","abstract_excerpt":"Distributed LLM serving is costly and often underutilizes hardware accelerators due to three key challenges: bubbles in pipeline-parallel deployments caused by the bimodal latency of prompt and token processing, GPU memory overprovisioning, and long recovery times in case of failures. In this paper, we propose D\\'ej\\`aVu, a system to address all these challenges using a versatile and efficient KV cache streaming library (D\\'ej\\`aVuLib). Using D\\'ej\\`aVuLib, we propose and implement efficient prompt-token disaggregation to reduce pipeline bubbles, microbatch swapping for efficient GPU memory ma"},"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":"2403.01876","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-03-04T09:32:05Z","cross_cats_sorted":[],"title_canon_sha256":"a18ea48afb22b447fb0e0f0aa2e558424ee97ddd4eae33c4edf0da655b10000f","abstract_canon_sha256":"3e0ae57aaf335501062feefc9cc9a17ae256e08d06f1ee360bcf61a9cdbd6f77"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:51:53.386892Z","signature_b64":"PRUPAXgYxuLHyVFTWXV8fNfUfWeoq4aCks5Tgr2B9EKlhhT/ytRvooYl2wk+HykxjvNPDpc5pU/xjIpA56BoDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c708e472b2e2381f8c74613a8359963e88561ca32dabbebca9f4fa6e47055d49","last_reissued_at":"2026-07-05T07:51:53.386375Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:51:53.386375Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"D\\'ej\\`aVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Amar Phanishayee, Ana Klimovic, Foteini Strati, Jakub Tarnawski, Sara McAllister","submitted_at":"2024-03-04T09:32:05Z","abstract_excerpt":"Distributed LLM serving is costly and often underutilizes hardware accelerators due to three key challenges: bubbles in pipeline-parallel deployments caused by the bimodal latency of prompt and token processing, GPU memory overprovisioning, and long recovery times in case of failures. In this paper, we propose D\\'ej\\`aVu, a system to address all these challenges using a versatile and efficient KV cache streaming library (D\\'ej\\`aVuLib). Using D\\'ej\\`aVuLib, we propose and implement efficient prompt-token disaggregation to reduce pipeline bubbles, microbatch swapping for efficient GPU memory ma"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.01876","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/2403.01876/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":"2403.01876","created_at":"2026-07-05T07:51:53.386443+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.01876v1","created_at":"2026-07-05T07:51:53.386443+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.01876","created_at":"2026-07-05T07:51:53.386443+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y4EOI4VS4I4B","created_at":"2026-07-05T07:51:53.386443+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y4EOI4VS4I4B7DDU","created_at":"2026-07-05T07:51:53.386443+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y4EOI4VS","created_at":"2026-07-05T07:51:53.386443+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29708","citing_title":"Demystifying the Design Space and Best Practices for Heterogeneous LLM Inference and Serving","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29708","citing_title":"Demystifying the Design Space and Best Practices for Heterogeneous LLM Inference and Serving","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2","json":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2.json","graph_json":"https://pith.science/api/pith-number/Y4EOI4VS4I4B7DDUME5IGWMWH2/graph.json","events_json":"https://pith.science/api/pith-number/Y4EOI4VS4I4B7DDUME5IGWMWH2/events.json","paper":"https://pith.science/paper/Y4EOI4VS"},"agent_actions":{"view_html":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2","download_json":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2.json","view_paper":"https://pith.science/paper/Y4EOI4VS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.01876&json=true","fetch_graph":"https://pith.science/api/pith-number/Y4EOI4VS4I4B7DDUME5IGWMWH2/graph.json","fetch_events":"https://pith.science/api/pith-number/Y4EOI4VS4I4B7DDUME5IGWMWH2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2/action/storage_attestation","attest_author":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2/action/author_attestation","sign_citation":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2/action/citation_signature","submit_replication":"https://pith.science/pith/Y4EOI4VS4I4B7DDUME5IGWMWH2/action/replication_record"}},"created_at":"2026-07-05T07:51:53.386443+00:00","updated_at":"2026-07-05T07:51:53.386443+00:00"}