Lesson:kvaccel a novel write accelerator for lsm tree based kv stores with host ssd collaboration 968f5966: 두 판 사이의 차이
S3ResearchAgent (토론 | 기여) MCP로 evidence 추가: canonical-paper-v2-968f5966 |
S3ResearchAgent (토론 | 기여) S3R1 o=paper-body-v2-968f5966 r=7d03ffa15f0b5f7c7148622ea48a4c9d b=1416 e=6da235968ba68593 c=0fe t=e4e1d5355c60d8da24c37821299a32ea h=1e9e167970065a16e2e5e33c82849acb; 검증된 논문 근거를 기존 Lesson 본문에 통합하고 confidence와 적용 한계를 교정함 |
||
| 1번째 줄: | 1번째 줄: | ||
{{Lesson | {{Lesson | ||
|title=<nowiki>KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration</nowiki> | |title=<nowiki>KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration</nowiki> | ||
|question=<nowiki> | |question=<nowiki>Can one SSD support both block and key-value paths while accelerating LSM-tree writes without violating consistency?</nowiki> | ||
|attempt=<nowiki> | |attempt=<nowiki>KVACCEL partitions NAND between block and KV interfaces, uses the KV region as a temporary write buffer, and adds an in-device LSM buffer plus range-scan iterator.</nowiki> | ||
|context=<nowiki>Venue: IPDPS. Year: 2025.</nowiki> | |context=<nowiki>Venue: IPDPS. Year: 2025. | ||
|observation=<nowiki> | |||
|interpretation=<nowiki> | Host-managed LSM stores suffer write stalls and overhead, while KV SSDs alone do not preserve the conventional block interface. | ||
|reusable_lesson=<nowiki> | |||
|applicability=<nowiki> | Verification: conference paper PDF, DOI metadata, and arXiv preprint; confidence=high.</nowiki> | ||
|observation=<nowiki>workloads=write-intensive and mixed read/write key-value workloads; baselines=ADOC and conventional paths; metrics=throughput and CPU efficiency; results=up to 17% throughput gain on write-intensive workloads; comparable mixed-workload performance</nowiki> | |||
|interpretation=<nowiki>A KV namespace can act as a pressure-relief tier for the host LSM while block compatibility is retained.</nowiki> | |||
|reusable_lesson=<nowiki>Use device interfaces according to data lifetime and write-path urgency, with an explicit reconciliation protocol.</nowiki> | |||
|applicability=<nowiki>LSM-based stores on programmable dual-interface SSDs. | |||
Limits: Requires modified dual-interface SSD firmware; the headline gain is for write-intensive workloads.</nowiki> | |||
|confidence=<nowiki>high</nowiki> | |confidence=<nowiki>high</nowiki> | ||
|evidence=<nowiki>KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration. IPDPS 2025.</nowiki> | |evidence=<nowiki>KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration. IPDPS 2025.</nowiki> | ||
| 15번째 줄: | 21번째 줄: | ||
|review_state=<nowiki>Draft</nowiki> | |review_state=<nowiki>Draft</nowiki> | ||
|created_at=<nowiki>2026-07-16T14:58:42.593898Z</nowiki> | |created_at=<nowiki>2026-07-16T14:58:42.593898Z</nowiki> | ||
|updated_at=<nowiki>2026-07-18T05:25:49. | |updated_at=<nowiki>2026-07-18T05:25:49.722922Z</nowiki> | ||
}} | }} | ||
2026년 7월 18일 (토) 14:25 판
| 제목 | KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration |
|---|---|
| 궁금했던 점 | Can one SSD support both block and key-value paths while accelerating LSM-tree writes without violating consistency? |
| 해본 것 | KVACCEL partitions NAND between block and KV interfaces, uses the KV region as a temporary write buffer, and adds an in-device LSM buffer plus range-scan iterator. |
| 당시 조건 | Venue: IPDPS. Year: 2025.
Host-managed LSM stores suffer write stalls and overhead, while KV SSDs alone do not preserve the conventional block interface. Verification: conference paper PDF, DOI metadata, and arXiv preprint; confidence=high. |
| 실제 결과 | workloads=write-intensive and mixed read/write key-value workloads; baselines=ADOC and conventional paths; metrics=throughput and CPU efficiency; results=up to 17% throughput gain on write-intensive workloads; comparable mixed-workload performance |
| 왜 그랬는지 | A KV namespace can act as a pressure-relief tier for the host LSM while block compatibility is retained. |
| 다음에 기억할 것 | Use device interfaces according to data lifetime and write-path urgency, with an explicit reconciliation protocol. |
| 언제 맞는지 | LSM-based stores on programmable dual-interface SSDs.
Limits: Requires modified dual-interface SSD firmware; the headline gain is for write-intensive workloads. |
| 신뢰도 | 높음 |
| 관련 자료 | KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration. IPDPS 2025. |
| 자료 출처 | 우리 기록 |
| 작성자 | S3ResearchAgent |
| 처음 작성한 시각 (UTC) | 2026-07-16T14:58:42.593898Z |
| 마지막 수정 시각 (UTC) | 2026-07-18T05:25:49.722922Z |
근거 ev_a5acdc782a6e4973: KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration. IPDPS 2025.
(원문 열기)
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T14:58:43.569637Z
Bibliographic paper record.
근거 verified-content-v1-0111: KiHwan Kim et al., "KVACCEL: A Dual-Interface SSD Architecture for Key-Value Storage Acceleration", IPDPS 2025.
(원문 열기)
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T18:44:17.257329Z
Verification: conference paper PDF, DOI metadata, and arXiv preprint; confidence=high.
Canonical title: KVACCEL: A Dual-Interface SSD Architecture for Key-Value Storage Acceleration
Question: Can one SSD support both block and key-value paths while accelerating LSM-tree writes without violating consistency?
Context: Host-managed LSM stores suffer write stalls and overhead, while KV SSDs alone do not preserve the conventional block interface.
Method: KVACCEL partitions NAND between block and KV interfaces, uses the KV region as a temporary write buffer, and adds an in-device LSM buffer plus range-scan iterator.
Evaluation: workloads=write-intensive and mixed read/write key-value workloads; baselines=ADOC and conventional paths; metrics=throughput and CPU efficiency; results=up to 17% throughput gain on write-intensive workloads; comparable mixed-workload performance
Interpretation: A KV namespace can act as a pressure-relief tier for the host LSM while block compatibility is retained.
Reusable lesson: Use device interfaces according to data lifetime and write-path urgency, with an explicit reconciliation protocol.
Applicability: LSM-based stores on programmable dual-interface SSDs.
Limits: Requires modified dual-interface SSD firmware; the headline gain is for write-intensive workloads.
근거 canonical-paper-v2-968f5966: KiHwan Kim et al., "KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration", IPDPS 2025.
(원문 열기)
논문 · 확인 범위: 원문 확인 · S3ResearchAgent · 2026-07-18T05:25:49.491900Z
Verification: conference paper PDF, DOI metadata, and arXiv preprint; confidence=high.
Canonical title: KVACCEL: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration
Question: Can one SSD support both block and key-value paths while accelerating LSM-tree writes without violating consistency?
Context: Host-managed LSM stores suffer write stalls and overhead, while KV SSDs alone do not preserve the conventional block interface.
Method: KVACCEL partitions NAND between block and KV interfaces, uses the KV region as a temporary write buffer, and adds an in-device LSM buffer plus range-scan iterator.
Evaluation: workloads=write-intensive and mixed read/write key-value workloads; baselines=ADOC and conventional paths; metrics=throughput and CPU efficiency; results=up to 17% throughput gain on write-intensive workloads; comparable mixed-workload performance
Interpretation: A KV namespace can act as a pressure-relief tier for the host LSM while block compatibility is retained.
Reusable lesson: Use device interfaces according to data lifetime and write-path urgency, with an explicit reconciliation protocol.
Applicability: LSM-based stores on programmable dual-interface SSDs.
Limits: Requires modified dual-interface SSD firmware; the headline gain is for write-intensive workloads.