{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DG737GQHELQO5GMF3CLUMPIBT2","short_pith_number":"pith:DG737GQH","schema_version":"1.0","canonical_sha256":"19bfbf9a0722e0ee9985d897463d019e9abe696e23ed4c29ee3f26da183b2f83","source":{"kind":"arxiv","id":"2508.00234","version":1},"attestation_state":"computed","paper":{"title":"Quality-of-Service Aware LLM Routing for Edge Computing with Multiple Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC","cs.MA"],"primary_cat":"cs.NI","authors_text":"Deke Guo, Jin Yang, Qiong Wu, Xu Chen, Zhiying Feng, Zhi Zhou","submitted_at":"2025-08-01T00:45:15Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities, leading to a significant increase in user demand for LLM services. However, cloud-based LLM services often suffer from high latency, unstable responsiveness, and privacy concerns. Therefore, multiple LLMs are usually deployed at the network edge to boost real-time responsiveness and protect data privacy, particularly for many emerging smart mobile and IoT applications. Given the varying response quality and latency of LLM services, a critical issue is how to route user requests from mobile and IoT devices to an appropriate"},"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":"2508.00234","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NI","submitted_at":"2025-08-01T00:45:15Z","cross_cats_sorted":["cs.AI","cs.DC","cs.MA"],"title_canon_sha256":"51f42e4c98e255358c5fa2fac8d5cc3d881bbefac2d4858353d8f3fba982bc28","abstract_canon_sha256":"dd8fc36f32b150fc089e5589c848a90d3a8e3e47b437e58fa09fb4f25ecfbdf1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:46:51.719955Z","signature_b64":"TVDYgxQr1gIZyMzHcqedM1El9WnNVGCYhxnmbei0KNjDyABK1UaKEEZL1TAQ32O4m5WBVplfxhjt5ocqXIkeCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"19bfbf9a0722e0ee9985d897463d019e9abe696e23ed4c29ee3f26da183b2f83","last_reissued_at":"2026-07-05T11:46:51.719528Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:46:51.719528Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quality-of-Service Aware LLM Routing for Edge Computing with Multiple Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC","cs.MA"],"primary_cat":"cs.NI","authors_text":"Deke Guo, Jin Yang, Qiong Wu, Xu Chen, Zhiying Feng, Zhi Zhou","submitted_at":"2025-08-01T00:45:15Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities, leading to a significant increase in user demand for LLM services. However, cloud-based LLM services often suffer from high latency, unstable responsiveness, and privacy concerns. Therefore, multiple LLMs are usually deployed at the network edge to boost real-time responsiveness and protect data privacy, particularly for many emerging smart mobile and IoT applications. Given the varying response quality and latency of LLM services, a critical issue is how to route user requests from mobile and IoT devices to an appropriate"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.00234","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/2508.00234/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":"2508.00234","created_at":"2026-07-05T11:46:51.719584+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.00234v1","created_at":"2026-07-05T11:46:51.719584+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.00234","created_at":"2026-07-05T11:46:51.719584+00:00"},{"alias_kind":"pith_short_12","alias_value":"DG737GQHELQO","created_at":"2026-07-05T11:46:51.719584+00:00"},{"alias_kind":"pith_short_16","alias_value":"DG737GQHELQO5GMF","created_at":"2026-07-05T11:46:51.719584+00:00"},{"alias_kind":"pith_short_8","alias_value":"DG737GQH","created_at":"2026-07-05T11:46:51.719584+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/DG737GQHELQO5GMF3CLUMPIBT2","json":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2.json","graph_json":"https://pith.science/api/pith-number/DG737GQHELQO5GMF3CLUMPIBT2/graph.json","events_json":"https://pith.science/api/pith-number/DG737GQHELQO5GMF3CLUMPIBT2/events.json","paper":"https://pith.science/paper/DG737GQH"},"agent_actions":{"view_html":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2","download_json":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2.json","view_paper":"https://pith.science/paper/DG737GQH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.00234&json=true","fetch_graph":"https://pith.science/api/pith-number/DG737GQHELQO5GMF3CLUMPIBT2/graph.json","fetch_events":"https://pith.science/api/pith-number/DG737GQHELQO5GMF3CLUMPIBT2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2/action/storage_attestation","attest_author":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2/action/author_attestation","sign_citation":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2/action/citation_signature","submit_replication":"https://pith.science/pith/DG737GQHELQO5GMF3CLUMPIBT2/action/replication_record"}},"created_at":"2026-07-05T11:46:51.719584+00:00","updated_at":"2026-07-05T11:46:51.719584+00:00"}