]> 2026-09-08T17:55:55+09:00 Lesson:telescope telemetry for gargantuan memory footprint applications 6a766ffb 3000 ko Tiering and telemetry for very large-memory applications. Limits: Effectiveness depends on page-table locality, TLB/page-walk behavior, and genuinely huge footprints. Telescope profiles upper page-table levels to identify coarse hot/cold subtrees and selectively refine them. S3ResearchAgent medium Venue: USENIX ATC. Year: 2024. Per-page sampling becomes too expensive and slow as memory footprints grow. Verification: official_abstract; confidence=high. 2026-07-16T14:56:51Z 2461238.1228125 Telescope: Telemetry for Gargantuan Memory Footprint Applications. USENIX ATC 2024. Hierarchical address-translation metadata provides scalable aggregation before expensive fine-grained inspection. Telescope: Telemetry for Gargantuan Memory Footprint Applications workloads=5 TB microbenchmark; 1–2 TB real benchmarks; baselines=state-of-the-art memory telemetry; metrics=precision; recall; CPU overhead; throughput; results=>90% precision and recall at 0.9% of one CPU; 5.6–34% throughput improvement How can systems identify hot and cold memory regions at terabyte-to-petabyte scale with little CPU cost? lab Use hierarchy to prune monitoring work at extreme scale. Draft 2026-07-18T15:00:41Z 2461240.1254745 2026-07-18T15:00:41Z 2461240.1254745 telescope telemetry for gargantuan memory footprint applications 6a766ffb 3000 ev_a18c475395544e46 2026-07-16T14:56:52Z 2461238.1228241 S3ResearchAgent Telescope: Telemetry for Gargantuan Memory Footprint Applications. USENIX ATC 2024. paper Bibliographic paper record. not_recorded telescope telemetry for gargantuan memory footprint applications 6a766ffb# ed619b35adaa6f1e4bd15b57da4a1d51 3000 verified-content-v1-0067 2026-07-16T18:43:02Z 2461238.2798843 S3ResearchAgent Alan Nair et al., "Telescope: Telemetry for Gargantuan Memory Footprint Applications", USENIX ATC 2024. paper Verification: official_abstract; confidence=high. Question: How can systems identify hot and cold memory regions at terabyte-to-petabyte scale with little CPU cost? Context: Per-page sampling becomes too expensive and slow as memory footprints grow. Method: Telescope profiles upper page-table levels to identify coarse hot/cold subtrees and selectively refine them. Evaluation: workloads=5 TB microbenchmark; 1–2 TB real benchmarks; baselines=state-of-the-art memory telemetry; metrics=precision; recall; CPU overhead; throughput; results=>90% precision and recall at 0.9% of one CPU; 5.6–34% throughput improvement Interpretation: Hierarchical address-translation metadata provides scalable aggregation before expensive fine-grained inspection. Reusable lesson: Use hierarchy to prune monitoring work at extreme scale. Applicability: Tiering and telemetry for very large-memory applications. Limits: Effectiveness depends on page-table locality, TLB/page-walk behavior, and genuinely huge footprints. not_recorded telescope telemetry for gargantuan memory footprint applications 6a766ffb# 38d7d0476b202d63fc4528de16e4c3d3 3000 canonical-paper-v2-6a766ffb 2026-07-18T05:39:20Z 2461239.7356482 S3ResearchAgent Alan Nair et al., "Telescope: Telemetry for Gargantuan Memory Footprint Applications", USENIX ATC 2024. paper Verification: official_abstract; confidence=medium. Question: How can systems identify hot and cold memory regions at terabyte-to-petabyte scale with little CPU cost? Context: Per-page sampling becomes too expensive and slow as memory footprints grow. Method: Telescope profiles upper page-table levels to identify coarse hot/cold subtrees and selectively refine them. Evaluation: workloads=5 TB microbenchmark; 1–2 TB real benchmarks; baselines=state-of-the-art memory telemetry; metrics=precision; recall; CPU overhead; throughput; results=>90% precision and recall at 0.9% of one CPU; 5.6–34% throughput improvement Interpretation: Hierarchical address-translation metadata provides scalable aggregation before expensive fine-grained inspection. Reusable lesson: Use hierarchy to prune monitoring work at extreme scale. Applicability: Tiering and telemetry for very large-memory applications. Limits: Effectiveness depends on page-table locality, TLB/page-walk behavior, and genuinely huge footprints. official_abstract telescope telemetry for gargantuan memory footprint applications 6a766ffb# de3eb590682f08e227a23fe3b53482a3 3000 2026-07-18T15:00:41Z 2461240.1254745 S3ResearchAgent verify_10aaa3fde1ac8bf47baf partial_source context 보존 원문 객체를 확보했으나 이 일괄 검증에서는 claim-bearing 범위를 재판정하지 않아 결론을 보류함. inconclusive 5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 보존 파일 objects/sha256/5b/5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 ev_a18c475395544e46 1484adc0c7952432e308be3de5079bda604f1f900d1c2352178fb105cf6f51ce telescope telemetry for gargantuan memory footprint applications 6a766ffb# 2e0fea5224906a46b8fb1ca1d0d7b2b6 3000 2026-07-18T15:00:41Z 2461240.1254745 S3ResearchAgent verify_6647644c1481efc8b836 partial_source context 보존 원문 객체를 확보했으나 이 일괄 검증에서는 claim-bearing 범위를 재판정하지 않아 결론을 보류함. inconclusive 5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 보존 파일 objects/sha256/5b/5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 verified-content-v1-0067 0cb7ea640098f4130ae47caf4d625e5c36e3d83b10624e7a02e4dbe0d35721d4 telescope telemetry for gargantuan memory footprint applications 6a766ffb# c694f14628b388c4199e7bad0d520867 3000 2026-07-18T15:00:41Z 2461240.1254745 S3ResearchAgent verify_6f11e4d59fefaec15205 partial_source context 보존 원문 객체를 확보했으나 이 일괄 검증에서는 claim-bearing 범위를 재판정하지 않아 결론을 보류함. inconclusive 5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 보존 파일 objects/sha256/5b/5ba2cea51fdfff24dd1684cad2e0bf26c7cbc7e8bd5a00805ba7fc21468a84d3 canonical-paper-v2-6a766ffb 2597ef6119a9657c8c5baefa45cbe9e7292eaae503c333258502bef6d967fe62 telescope telemetry for gargantuan memory footprint applications 6a766ffb# aac5c6e917bbfb6ef7374f3453f91636