doc_id stringlengths 11 42 | source_type stringclasses 5
values | source_url stringlengths 32 176 | license stringclasses 8
values | title stringlengths 3 654 | content stringlengths 439 232k | project stringclasses 35
values | tags listlengths 2 31 | word_count int32 71 33.5k | lang stringclasses 1
value | quality_score float64 0.35 0.9 | scraped_at timestamp[us, tz=UTC]date 2026-09-21 04:26:58 2026-09-21 12:13:36 |
|---|---|---|---|---|---|---|---|---|---|---|---|
arxiv-2509.22351-11 | arxiv | https://arxiv.org/abs/2509.22351 | CC-BY-4.0 | I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 3 Results › 3.3 ETL execution | Title: I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 3 Results › 3.3 ETL execution
arXiv: 2509.22351
### 3.3 ETL execution
When run on the initial datasets from the two hospitals considered in our scenario, two databases $H_{1}$ and $H_{2}$ are consolidated – see Table 8 ... | papers | [
"paper",
"arxiv_id:2509.22351",
"relevance:4",
"chunk:11"
] | 137 | en | 0.35 | 2026-09-21T09:48:39.773413Z |
arxiv-2509.22351-12 | arxiv | https://arxiv.org/abs/2509.22351 | CC-BY-4.0 | I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 3 Results › 3.4 Interoperability assessment | Title: I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 3 Results › 3.4 Interoperability assessment
arXiv: 2509.22351
### 3.4 Interoperability assessment
Finally, I-ETL reports on the overall interoperability of the target database by computing and displaying our set of metr... | papers | [
"paper",
"arxiv_id:2509.22351",
"relevance:4",
"chunk:12"
] | 814 | en | 0.475 | 2026-09-21T09:48:39.773416Z |
arxiv-2509.22351-13 | arxiv | https://arxiv.org/abs/2509.22351 | CC-BY-4.0 | I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 4 Discussion | Title: I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 4 Discussion
arXiv: 2509.22351
## 4 Discussion
### 4.1 Challenges and limitations
The primary challenge we faced was to design a conceptual data model that could fit the various kinds of data brought by hospitals and c... | papers | [
"paper",
"arxiv_id:2509.22351",
"relevance:4",
"chunk:13"
] | 1,328 | en | 0.5 | 2026-09-21T09:48:39.773419Z |
arxiv-2509.22351-14 | arxiv | https://arxiv.org/abs/2509.22351 | CC-BY-4.0 | I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 5 Conclusions | Title: I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — 5 Conclusions
arXiv: 2509.22351
## 5 Conclusions
In this paper, we presented I-ETL, a framework for integrating heterogeneous healthcare datasets with interoperability as a first-class citizen. Our contributions are the... | papers | [
"paper",
"arxiv_id:2509.22351",
"relevance:4",
"chunk:14"
] | 534 | en | 0.525 | 2026-09-21T09:48:39.773423Z |
arxiv-2509.22351-15 | arxiv | https://arxiv.org/abs/2509.22351 | CC-BY-4.0 | I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — Declarations | Title: I-ETL: an interoperability-aware health (meta) data pipeline to enable federated analyses — Declarations
arXiv: 2509.22351
## Declarations
#### Ethics approval and consent to participate
Not applicable.
#### Consent for publication
Not applicable.
#### Data availability
The datasets used in this manuscrip... | papers | [
"paper",
"arxiv_id:2509.22351",
"relevance:4",
"chunk:15"
] | 154 | en | 0.35 | 2026-09-21T09:48:39.773425Z |
arxiv-2601.19176-00 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 1 Introduction | Title: Create Benchmarks for Data Lakes
arXiv: 2601.19176 | License: CC-BY-4.0
Abstract: Data lakes have emerged as a flexible and scalable solution for storing and analyzing large volumes of heterogeneous data, including structured, semi-structured, and unstructured formats. Despite their growing adoption in both ind... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:0"
] | 1,067 | en | 0.425 | 2026-09-21T09:48:18.737046Z |
arxiv-2601.19176-01 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 2 Related Work | Title: Create Benchmarks for Data Lakes — 2 Related Work
arXiv: 2601.19176
## 2 Related Work
Currently, there are still no standardized benchmarks to benchmark the performance of data lakes. TPC-H[5] and TPC-DS[6] issued by the Transaction Processing Performance Council are still widely adopted by the big data system... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:1"
] | 533 | en | 0.525 | 2026-09-21T09:48:18.737056Z |
arxiv-2601.19176-02 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 3 Approach | Title: Create Benchmarks for Data Lakes — 3 Approach
arXiv: 2601.19176
## 3 Approach
Our approach to solving the problem is to design a standardized data lake benchmark to compare different data lake frameworks together.
For the source dataset, we are considering using structured data, unstructured data, and semi-st... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:2"
] | 212 | en | 0.4 | 2026-09-21T09:48:18.737061Z |
arxiv-2601.19176-03 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 4 Design | Title: Create Benchmarks for Data Lakes — 4 Design
arXiv: 2601.19176
## 4 Design
The system architecture of our benchmark framework is listed below.
The benchmark system starts by downloading the raw dataset. After much effort, we collected data from IMDb (Internet Movie Database)—a movie rating CSV dataset (structu... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:3"
] | 884 | en | 0.4 | 2026-09-21T09:48:18.737065Z |
arxiv-2601.19176-04 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 5 Evaluation | Title: Create Benchmarks for Data Lakes — 5 Evaluation
arXiv: 2601.19176
## 5 Evaluation
### 5.1 Experiment Setup
For evaluation, we are using a MacBook Pro with 16GB memory and 512 GB disk. The chip is an M1 Pro and the OS is macOS Monterey Version 12.1.
Brand | Apple |
---|---|
OS | macOS Monterey (12.1) |
Memo... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:4"
] | 638 | en | 0.4 | 2026-09-21T09:48:18.737069Z |
arxiv-2601.19176-05 | arxiv | https://arxiv.org/abs/2601.19176 | CC-BY-4.0 | Create Benchmarks for Data Lakes — 6 Conclusion & Future works | Title: Create Benchmarks for Data Lakes — 6 Conclusion & Future works
arXiv: 2601.19176
## 6 Conclusion & Future works
In this paper, we presented a new benchmark for evaluating data lakes that can store all types of datasets including structured, semi-structured, and unstructured datasets. We described the methods t... | papers | [
"paper",
"arxiv_id:2601.19176",
"relevance:4",
"chunk:5"
] | 410 | en | 0.4 | 2026-09-21T09:48:18.737072Z |
arxiv-2605.09994-00 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training
arXiv: 2605.09994 | License: CC-BY-4.0
Abstract: Modern Large Foundation Model (LFM) training has transformed the data pipeline from a static ingestion layer into a dynamic component that must co-evolve with the training... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:0"
] | 444 | en | 0.525 | 2026-09-21T09:47:38.045804Z |
arxiv-2605.09994-01 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 1. Introduction | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 1. Introduction
arXiv: 2605.09994
## 1. Introduction
Data loading has long been a well-studied component of model training. In conventional deep learning workloads, the data pipeline is straightforward: a static, pre-... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:1"
] | 1,251 | en | 0.625 | 2026-09-21T09:47:38.045823Z |
arxiv-2605.09994-02 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 2. Background | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 2. Background
arXiv: 2605.09994
## 2. Background
### 2.1. Requirements for LFM Training
SPMD execution and the Global Batch.
Modern LFM training follows the Single Program Multiple Data (SPMD) execution
model, where ... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:2"
] | 1,476 | en | 0.475 | 2026-09-21T09:47:38.045828Z |
arxiv-2605.09994-03 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 3. BatchWeave Overview | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 3. BatchWeave Overview
arXiv: 2605.09994
## 3. BatchWeave Overview
Object storage provides durable, decentralized, immutable object writes, but offers no mechanism for expressing relationships between objects, no way ... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:3"
] | 645 | en | 0.5 | 2026-09-21T09:47:38.045832Z |
arxiv-2605.09994-04 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 4. Transactional Global Batch | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 4. Transactional Global Batch
arXiv: 2605.09994
## 4. Transactional Global Batch
The Transactional Global Batch (TGB) abstraction gives each $\mathcal{B}_{s}$ a durable, atomically visible representation on object sto... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:4"
] | 1,092 | en | 0.45 | 2026-09-21T09:47:38.045836Z |
arxiv-2605.09994-05 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 5. Manifest-Based Control Plane | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 5. Manifest-Based Control Plane
arXiv: 2605.09994
## 5. Manifest-Based Control Plane
The manifest-based control plane of BatchWeave manages three responsibilities: linearizing the global batch sequence across concurre... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:5"
] | 1,455 | en | 0.65 | 2026-09-21T09:47:38.045839Z |
arxiv-2605.09994-06 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 6. Implementation | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 6. Implementation
arXiv: 2605.09994
## 6. Implementation
BatchWeave is implemented as a Python SDK with a Rust core. The Rust layer handles latency-critical execution: multi-threaded object prefetching, manifest polli... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:6"
] | 229 | en | 0.425 | 2026-09-21T09:47:38.045843Z |
arxiv-2605.09994-07 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.1. Experimental Setup | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.1. Experimental Setup
arXiv: 2605.09994
### 7.1. Experimental Setup
Infrastructure. All experiments run on a shared kjob cluster backed by Baidu Object Storage (BOS) (Li et al., 2025). Each node has ... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:7"
] | 783 | en | 0.525 | 2026-09-21T09:47:38.045846Z |
arxiv-2605.09994-08 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.2. End-to-End Training Performance | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.2. End-to-End Training Performance
arXiv: 2605.09994
### 7.2. End-to-End Training Performance
Unless otherwise noted, BatchWeave uses dedicated CPU producer nodes and no train-side producers: 32 node... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:8"
] | 396 | en | 0.475 | 2026-09-21T09:47:38.045850Z |
arxiv-2605.09994-09 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.3. Producer Scalability and DAC | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.3. Producer Scalability and DAC
arXiv: 2605.09994
### 7.3. Producer Scalability and DAC
Producer throughput scaling. Figure 6 sweeps producer count and payload size, comparing BatchWeave against Kafk... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:9"
] | 420 | en | 0.525 | 2026-09-21T09:47:38.045853Z |
arxiv-2605.09994-10 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.4. Consumer Efficiency and Read Amplification | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.4. Consumer Efficiency and Read Amplification
arXiv: 2605.09994
### 7.4. Consumer Efficiency and Read Amplification
Each rank issues a single range read targeting only its own data slice, so BatchWea... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:10"
] | 238 | en | 0.425 | 2026-09-21T09:47:38.045856Z |
arxiv-2605.09994-11 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.5. Checkpoint-Driven Lifecycle Management | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 7. Evaluation › 7.5. Checkpoint-Driven Lifecycle Management
arXiv: 2605.09994
### 7.5. Checkpoint-Driven Lifecycle Management
Figure 9 compares two otherwise identical 1,010-step runs with checkpoints every 10 steps a... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:11"
] | 154 | en | 0.35 | 2026-09-21T09:47:38.045859Z |
arxiv-2605.09994-12 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 8. Related Work | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 8. Related Work
arXiv: 2605.09994
## 8. Related Work
### 8.1. Message Queues and Streaming Systems
Distributed message queues such as Apache Kafka (Kreps et al., 2011) and Apache Pulsar (The Apache Software Foundatio... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:12"
] | 526 | en | 0.65 | 2026-09-21T09:47:38.045863Z |
arxiv-2605.09994-13 | arxiv | https://arxiv.org/abs/2605.09994 | CC-BY-4.0 | BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 9. Conclusion | Title: BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training — 9. Conclusion
arXiv: 2605.09994
## 9. Conclusion
Modern LFM training demands more from the data pipeline than record delivery: the Global Batch is the unit of distributed optimization, and its boundaries, ordering, an... | papers | [
"paper",
"arxiv_id:2605.09994",
"relevance:5",
"chunk:13"
] | 122 | en | 0.425 | 2026-09-21T09:47:38.045867Z |
arxiv-2605.27571-00 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems
arXiv: 2605.27571 | License: CC-BY-4.0
Abstract: Modern analytics systems are fundamentally reactive, requiring users to define queries over increasingly complex and continuously evolving data. In real-time streaming environments, this p... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:0"
] | 318 | en | 0.5 | 2026-09-21T09:47:28.988843Z |
arxiv-2605.27571-01 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 1. Introduction | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 1. Introduction
arXiv: 2605.27571
## 1. Introduction
Data analytics systems have evolved from manual SQL workflows to dashboards and, more recently, LLM-powered copilots. Despite these advances, the dominant interaction model remains ... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:1"
] | 442 | en | 0.475 | 2026-09-21T09:47:28.988862Z |
arxiv-2605.27571-02 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 2. Related Work | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 2. Related Work
arXiv: 2605.27571
## 2. Related Work
LLM-powered data agents are emerging as a general abstraction for automating data management, preparation, analysis, visualization, and reporting. Recent surveys characterize these ... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:2"
] | 293 | en | 0.425 | 2026-09-21T09:47:28.988866Z |
arxiv-2605.27571-03 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 3. Architecture: Discovery Agents over Streaming Data | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 3. Architecture: Discovery Agents over Streaming Data
arXiv: 2605.27571
## 3. Architecture: Discovery Agents over Streaming Data
The proposed system implements a multi-agent architecture for transforming real-time data streams into de... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:3"
] | 241 | en | 0.525 | 2026-09-21T09:47:28.988870Z |
arxiv-2605.27571-04 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 4. Agent Semantics and Artifact Contracts | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 4. Agent Semantics and Artifact Contracts
arXiv: 2605.27571
## 4. Agent Semantics and Artifact Contracts
The system decomposes the analytics lifecycle into specialized agents, each responsible for a semantic transformation over typed ... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:4"
] | 1,061 | en | 0.65 | 2026-09-21T09:47:28.988873Z |
arxiv-2605.27571-05 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 5. Use Cases and Early Lessons | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 5. Use Cases and Early Lessons
arXiv: 2605.27571
## 5. Use Cases and Early Lessons
Given the exploratory nature of the system, we evaluate it through representative use cases rather than a formal benchmark. The goal is to characterize... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:5"
] | 398 | en | 0.45 | 2026-09-21T09:47:28.988877Z |
arxiv-2605.27571-06 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 6. Position and Open Challenges | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 6. Position and Open Challenges
arXiv: 2605.27571
## 6. Position and Open Challenges
We argue that proactive insight discovery should become a first-class workload for agent-first data systems. The proposed architecture shifts analyti... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:6"
] | 247 | en | 0.525 | 2026-09-21T09:47:28.988880Z |
arxiv-2605.27571-07 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion
arXiv: 2605.27571
## 7. Conclusion
We presented a multi-agent architecture for autonomous discovery over real-time data streams. By combining Kafka-based coordination, Flink-based stream processing, and LLM-powered agent... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:7"
] | 145 | en | 0.375 | 2026-09-21T09:47:28.988883Z |
arxiv-2605.27571-08 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.1. Demo Summary | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.1. Demo Summary
arXiv: 2605.27571
### A.1. Demo Summary
The NYC Parks Events Listing dataset contains public event records across New York City parks. Each record includes temporal, geographic, textual, categorical, ... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:8"
] | 189 | en | 0.35 | 2026-09-21T09:47:28.988887Z |
arxiv-2605.27571-09 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.2. End-to-End Lineage Example | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.2. End-to-End Lineage Example
arXiv: 2605.27571
### A.2. End-to-End Lineage Example
To illustrate the full discovery pipeline, we trace a single high-priority hypothesis through all transformation stages. This exampl... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:9"
] | 720 | en | 0.625 | 2026-09-21T09:47:28.988891Z |
arxiv-2605.27571-11 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.4. Generated Application Structure | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.4. Generated Application Structure
arXiv: 2605.27571
### A.4. Generated Application Structure
The Deploy Agent packages the generated artifacts into a runnable application. The resulting directory contains the backen... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:11"
] | 154 | en | 0.4 | 2026-09-21T09:47:28.988897Z |
arxiv-2605.27571-12 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.5. Execution and Discovered Insights | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.5. Execution and Discovered Insights
arXiv: 2605.27571
### A.5. Execution and Discovered Insights
The application can be launched with:
```
docker-compose up --build
```
This starts the backend and frontend service... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:12"
] | 174 | en | 0.4 | 2026-09-21T09:47:28.988903Z |
arxiv-2605.27571-13 | arxiv | https://arxiv.org/abs/2605.27571 | CC-BY-4.0 | Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.6. Lessons from the Demo | Title: Discovery Agents for Real-Time Analytics: Toward Proactive Insight Systems — 7. Conclusion › A.6. Lessons from the Demo
arXiv: 2605.27571
### A.6. Lessons from the Demo
The NYC Parks Events demo highlights five practical lessons.
- (1)
Typed contracts improve reliability. Explicit intermediate artifacts make ... | papers | [
"paper",
"arxiv_id:2605.27571",
"relevance:6",
"chunk:13"
] | 165 | en | 0.375 | 2026-09-21T09:47:28.988909Z |
arxiv-2605.30308-00 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale
arXiv: 2605.30308 | License: CC-BY-4.0
Abstract: Modern table formats such as Apache Iceberg compute and store metadata-commit timestamps, record counts, and column-level statistics such as null counts and value bound... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:0"
] | 390 | en | 0.525 | 2026-09-21T09:48:51.649653Z |
arxiv-2605.30308-01 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 1. Introduction | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 1. Introduction
arXiv: 2605.30308
## 1. Introduction
Data quality monitoring is essential for data-intensive organizations, yet the dominant paradigm treats it as a read-time concern: tools such as dbt tests (6), G... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:1"
] | 317 | en | 0.575 | 2026-09-21T09:48:51.649664Z |
arxiv-2605.30308-02 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 2. Motivation | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 2. Motivation
arXiv: 2605.30308
## 2. Motivation
### 2.1. Available Metadata in Table Formats
Each data file registered in an Iceberg *manifest*—the metadata layer that tracks which files comprise a table snapshot... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:2"
] | 578 | en | 0.8 | 2026-09-21T09:48:51.649668Z |
arxiv-2605.30308-03 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 3. System Architecture | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 3. System Architecture
arXiv: 2605.30308
## 3. System Architecture
Writer identity. System operations such as compaction, sorting, and purging create new snapshots without changing underlying data. Without distingu... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:3"
] | 163 | en | 0.475 | 2026-09-21T09:48:51.649671Z |
arxiv-2605.30308-04 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 4. Production Experience | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 4. Production Experience
arXiv: 2605.30308
## 4. Production Experience
### 4.1. Resource Impact
The metadata-first approach is deployed across 200,000+ tables (800+ PB) at LinkedIn. Prior to this, scan-based profi... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:4"
] | 481 | en | 0.6 | 2026-09-21T09:48:51.649674Z |
arxiv-2605.30308-05 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 5. Proposed Extensions and Validation | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 5. Proposed Extensions and Validation
arXiv: 2605.30308
## 5. Proposed Extensions and Validation
Our experience suggests that table formats should treat data quality as a first-class design concern. Below we propos... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:5"
] | 1,176 | en | 0.8 | 2026-09-21T09:48:51.649678Z |
arxiv-2605.30308-06 | arxiv | https://arxiv.org/abs/2605.30308 | CC-BY-4.0 | Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 6. Related Work | Title: Zero-Scan Data Quality: Leveraging Table Format Metadata for Continuous Observability at Scale — 6. Related Work
arXiv: 2605.30308
## 6. Related Work
Data profiling has a long history (13). SPADE (12) synthesizes DQ assertions for LLM pipelines but remains query-based. Commercial formats like Delta Lake (4) an... | papers | [
"paper",
"arxiv_id:2605.30308",
"relevance:4",
"chunk:6"
] | 128 | en | 0.45 | 2026-09-21T09:48:51.649681Z |
arxiv-2605.31475-00 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Abstract | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases
arXiv: 2605.31475 | License: CC-BY-4.0
Abstract: DBLog is a change-data-capture (CDC) mechanism for copying a table or selected keys from a source database while continuing to stream new changes from its comm... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:0"
] | 616 | en | 0.725 | 2026-09-21T09:47:40.874274Z |
arxiv-2605.31475-01 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Introduction | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Introduction
arXiv: 2605.31475
## Introduction
DBLog, introduced in the Netflix Tech Blog in 2019 [dblog_netflix_blog]
and described in the 2020 DBLog paper [andreakis2020dblog], is a
change-data-capture (... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:1"
] | 949 | en | 0.8 | 2026-09-21T09:47:40.874284Z |
arxiv-2605.31475-02 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example
arXiv: 2605.31475
## Bootstrapping and Advancing a Virtual Cut with a Running Example
We begin with a concrete backfill before introducing t... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:2"
] | 116 | en | 0.35 | 2026-09-21T09:47:40.874288Z |
arxiv-2605.31475-03 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .1 Setup | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .1 Setup
arXiv: 2605.31475
### .1 Setup
The source database contains an account table
accounts(account_id, balance)
initially populated ... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:3"
] | 130 | en | 0.35 | 2026-09-21T09:47:40.874293Z |
arxiv-2605.31475-04 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .2 What the Source Log Records During the Backfill Window | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .2 What the Source Log Records During the Backfill Window
arXiv: 2605.31475
### .2 What the Source Log Records During the Backfill Window
... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:4"
] | 192 | en | 0.35 | 2026-09-21T09:47:40.874297Z |
arxiv-2605.31475-05 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .3 What the Chunks Read | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .3 What the Chunks Read
arXiv: 2605.31475
### .3 What the Chunks Read
DBLog reads the table while source writes continue. It partitions th... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:5"
] | 329 | en | 0.475 | 2026-09-21T09:47:40.874300Z |
arxiv-2605.31475-06 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .4 The Chosen Point of Equivalence | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .4 The Chosen Point of Equivalence
arXiv: 2605.31475
### .4 The Chosen Point of Equivalence
The backfill targets a specific source-log mom... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:6"
] | 121 | en | 0.35 | 2026-09-21T09:47:40.874303Z |
arxiv-2605.31475-07 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .5 How the Model Assembles the Replay | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .5 How the Model Assembles the Replay
arXiv: 2605.31475
### .5 How the Model Assembles the Replay
The model treats every observed source-l... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:7"
] | 685 | en | 0.475 | 2026-09-21T09:47:40.874306Z |
arxiv-2605.31475-08 | arxiv | https://arxiv.org/abs/2605.31475 | CC-BY-4.0 | A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .6 Advancing the Cut to a Later Frontier | Title: A Theoretical Study of DBLog: Certified Virtual Cuts for a Snapshot-Equivalent Replay of Live Databases — Bootstrapping and Advancing a Virtual Cut with a Running Example › .6 Advancing the Cut to a Later Frontier
arXiv: 2605.31475
### .6 Advancing the Cut to a Later Frontier
The seven-item replay above forms ... | papers | [
"paper",
"arxiv_id:2605.31475",
"relevance:5",
"chunk:8"
] | 397 | en | 0.475 | 2026-09-21T09:47:40.874309Z |
arxiv-2606.03464-00 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance
arXiv: 2606.03464 | License: CC-BY-4.0
Abstract: Data analytics increasingly runs on distributed lakehouse systems, where platform operators must optimise monetary, resource, and environmental costs. Query Performance... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:0"
] | 955 | en | 0.525 | 2026-09-21T09:49:16.303715Z |
arxiv-2606.03464-01 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — II Related Work | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — II Related Work
arXiv: 2606.03464
## II Related Work
In this section, we first review the work on runtime variance in distributed analytics. We then discuss runtime QPP for distributed analytics and low-carbon sche... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:1"
] | 920 | en | 0.525 | 2026-09-21T09:49:16.303725Z |
arxiv-2606.03464-02 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — III Experimental Design | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — III Experimental Design
arXiv: 2606.03464
## III Experimental Design
In this section, we outline the experimental setup, including the lakehouse architecture, cluster resources, workloads, and key configuration cho... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:2"
] | 769 | en | 0.475 | 2026-09-21T09:49:16.303729Z |
arxiv-2606.03464-03 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — IV Study 1: Variance Characterisation | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — IV Study 1: Variance Characterisation
arXiv: 2606.03464
## IV Study 1: Variance Characterisation
In this first study, we establish a macro view of query runtime variance in distributed lakehouses by examining how v... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:3"
] | 1,206 | en | 0.475 | 2026-09-21T09:49:16.303733Z |
arxiv-2606.03464-04 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — V Study 2: Factor Analysis | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — V Study 2: Factor Analysis
arXiv: 2606.03464
## V Study 2: Factor Analysis
To isolate the key contributors to runtime uncertainty, this section investigates major sources of variance in lakehouses across deployment... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:4"
] | 1,177 | en | 0.425 | 2026-09-21T09:49:16.303737Z |
arxiv-2606.03464-05 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VI Study 3: Impact on QPP | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VI Study 3: Impact on QPP
arXiv: 2606.03464
## VI Study 3: Impact on QPP
After establishing that runtime variance is a general challenge for lakehouses, this study examines how the variance regime of runtime labels... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:5"
] | 787 | en | 0.45 | 2026-09-21T09:49:16.303740Z |
arxiv-2606.03464-06 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VII Study 4: Low-Carbon Scheduling | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VII Study 4: Low-Carbon Scheduling
arXiv: 2606.03464
## VII Study 4: Low-Carbon Scheduling
This section presents low-carbon scheduling as a case study to examine how variance in lakehouse performance can impact pre... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:6"
] | 910 | en | 0.55 | 2026-09-21T09:49:16.303743Z |
arxiv-2606.03464-07 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VIII Discussion | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — VIII Discussion
arXiv: 2606.03464
## VIII Discussion
This section discusses the implications of addressing runtime variance and potential threats to validity.
### VIII-A Implications of the Results
Our investigat... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:7"
] | 601 | en | 0.45 | 2026-09-21T09:49:16.303747Z |
arxiv-2606.03464-08 | arxiv | https://arxiv.org/abs/2606.03464 | CC-BY-4.0 | Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — IX Conclusion | Title: Predicting Lakehouse Performance in Clouds: An Empirical Exploration of Query Runtime Variance — IX Conclusion
arXiv: 2606.03464
## IX Conclusion
In this paper, we investigated how runtime variance manifests in cloud deployments of distributed data lakehouses and how it propagates into query performance predic... | papers | [
"paper",
"arxiv_id:2606.03464",
"relevance:3",
"chunk:8"
] | 249 | en | 0.425 | 2026-09-21T09:49:16.303750Z |
arxiv-2608.00501-00 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log | Title: Machine-Checked Dual-Write Recovery from a Commit Log
arXiv: 2608.00501 | License: CC-BY-4.0
Abstract: Applications often need to make related facts durable in two independent systems without a transaction spanning both. If the process crashes after the second system accepts an operation but before a source-sid... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:0"
] | 902 | en | 0.55 | 2026-09-21T09:48:25.066320Z |
arxiv-2608.00501-01 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — Dual writes. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — Dual writes.
arXiv: 2608.00501
### Dual writes.
The incident is an example of a *dual write*. Related facts
become durable under independent authorities without one commit
spanning both. The source may be a database, while the second
authority may be a ma... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:1"
] | 213 | en | 0.475 | 2026-09-21T09:48:25.066333Z |
arxiv-2608.00501-02 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — Why formalize now. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — Why formalize now.
arXiv: 2608.00501
### Why formalize now.
The engineering guidance for this problem is established. Prior work recommends deriving state from a log, retrying failed deliveries, using stable identities, and fencing stale workers (Kleppman... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:2"
] | 211 | en | 0.45 | 2026-09-21T09:48:25.066336Z |
arxiv-2608.00501-03 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — Results. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — Results.
arXiv: 2608.00501
### Results.
The results are organized around three parts of the recovery problem.
*Source-side impossibility result.*
Sections 3 and 4 construct reachable
post-crash states that agree on the durable information available to
re... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:3"
] | 305 | en | 0.475 | 2026-09-21T09:48:25.066340Z |
arxiv-2608.00501-06 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Histories, frontiers, materialization. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Histories, frontiers, materialization.
arXiv: 2608.00501
### Histories, frontiers, materialization.
A *source event* $e$ on a key $k$ inserts or updates that key to a
value, or deletes it. A committed *source history* $H$ is a finite
sequen... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:6"
] | 396 | en | 0.4 | 2026-09-21T09:48:25.066349Z |
arxiv-2608.00501-07 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › The acceptance boundary. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › The acceptance boundary.
arXiv: 2608.00501
### The acceptance boundary.
The *effect machine* extends $\sigma$ with one field and one
register. A state is $t=\langle\sigma,E,\gamma\rangle$ : the core,
an *emissions ledger* $E$ , and a *gener... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:7"
] | 262 | en | 0.4 | 2026-09-21T09:48:25.066352Z |
arxiv-2608.00501-08 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Hazards and missing deliveries. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Hazards and missing deliveries.
arXiv: 2608.00501
### Hazards and missing deliveries.
Three predicates over $t$ classify how a ledger can be wrong, and one
checks for missing deliveries. The *obligations* at a frontier are the
scoped, front... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:8"
] | 301 | en | 0.45 | 2026-09-21T09:48:25.066355Z |
arxiv-2608.00501-09 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › What each side must provide. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › What each side must provide.
arXiv: 2608.00501
### What each side must provide.
The model abstracts over both ends. On the source side, the analysis assumes only a durable, ordered, append-only committed history: each fact either is in $H$ ... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:9"
] | 390 | en | 0.45 | 2026-09-21T09:48:25.066358Z |
arxiv-2608.00501-10 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Re-drives, policies, observations. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Re-drives, policies, observations.
arXiv: 2608.00501
### Re-drives, policies, observations.
A *re-drive policy* $P$ maps a state to a batch of payloads. The
re-drive relation $t\mathrel{\rightsquigarrow_{P\langle f\rangle}}t^{\prime}$ is on... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:10"
] | 245 | en | 0.425 | 2026-09-21T09:48:25.066361Z |
arxiv-2608.00501-11 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Mapping the model to systems. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 2. The Model › Mapping the model to systems.
arXiv: 2608.00501
### Mapping the model to systems.
Table 1 fixes the substitution between practitioner vocabulary and model objects. Two rows distinguish roles. An outbox table is the committed log when a rela... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:11"
] | 318 | en | 0.45 | 2026-09-21T09:48:25.066364Z |
arxiv-2608.00501-12 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 3. The Store-Side Observation Bound | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 3. The Store-Side Observation Bound
arXiv: 2608.00501
## 3. The Store-Side Observation Bound
After a crash, recovery must decide what to re-send. The first two theorems show that no policy computed only from the durable store and its control plane is corr... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:12"
] | 1,017 | en | 0.475 | 2026-09-21T09:48:25.066367Z |
arxiv-2608.00501-13 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 4. The Checkpoint Window | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 4. The Checkpoint Window
arXiv: 2608.00501
## 4. The Checkpoint Window
The observation bound uses two earlier reconciliation choices. To show the same ambiguity in a fixed implementation, we add a durable cursor and study a deterministic deliver-then-chec... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:13"
] | 1,243 | en | 0.4 | 2026-09-21T09:48:25.066370Z |
arxiv-2608.00501-14 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 5. Reading the Sink | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 5. Reading the Sink
arXiv: 2608.00501
## 5. Reading the Sink
The pairs in the last two sections differ in the sink-side accepted
record. A policy that reads this record can compute what is still
missing. The *sink delta* at frontier
$f$ is the committed o... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:14"
] | 830 | en | 0.55 | 2026-09-21T09:48:25.066373Z |
arxiv-2608.00501-16 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › The channel machine. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › The channel machine.
arXiv: 2608.00501
### The channel machine.
A channel state is $u=\langle t,W,A,\varphi\rangle$ : an inner
effect state $t$ , the *wire* $W$ (emissions sent and not yet
resolved), the sink's durabl... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:16"
] | 319 | en | 0.5 | 2026-09-21T09:48:25.066379Z |
arxiv-2608.00501-17 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › An old request can arrive after recovery. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › An old request can arrive after recovery.
arXiv: 2608.00501
### An old request can arrive after recovery.
Figure 3 shows an old send that remains in flight across the crash. Recovery reads the accepted record, sends t... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:17"
] | 452 | en | 0.45 | 2026-09-21T09:48:25.066382Z |
arxiv-2608.00501-18 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › The arrival fence. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › The arrival fence.
arXiv: 2608.00501
### The arrival fence.
The *fenced re-drive* at $f$ is one atomic act: heal the store,
compute $\Delta$ from the accepted record, land the batch in the sent
ledger and synchronousl... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:18"
] | 233 | en | 0.5 | 2026-09-21T09:48:25.066385Z |
arxiv-2608.00501-19 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Fence placement. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Fence placement.
arXiv: 2608.00501
### Fence placement.
The composite's shape is forced by a discipline the machine inherits:
re-drives stamp at the crash-time generation (the reconcile is
generation-constant, and res... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:19"
] | 216 | en | 0.425 | 2026-09-21T09:48:25.066388Z |
arxiv-2608.00501-20 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Cost of the arrival fence. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Cost of the arrival fence.
arXiv: 2608.00501
### Cost of the arrival fence.
The following corollary shows that fencing can reject a delayed delivery needed at a later frontier.
###### Corollary 6.0 (Rejection of late... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:20"
] | 276 | en | 0.45 | 2026-09-21T09:48:25.066390Z |
arxiv-2608.00501-21 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Validation controls. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 6. Stragglers and the Arrival Fence › Validation controls.
arXiv: 2608.00501
### Validation controls.
Two companion results locate the modeling choices. On a twin machine whose crash wipes the wire, the same schedule ends exactly-once with a duplicate-fre... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:21"
] | 117 | en | 0.35 | 2026-09-21T09:48:25.066393Z |
arxiv-2608.00501-22 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › The concurrency machine. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › The concurrency machine.
arXiv: 2608.00501
### The concurrency machine.
The machine carries
$w=\langle\sigma,J,S,A,\varphi,\Gamma,\mathit{hwm},\mathit{fl}\rangle$ . Its fields are the store, a *journal* $J$ (... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:22"
] | 308 | en | 0.45 | 2026-09-21T09:48:25.066396Z |
arxiv-2608.00501-23 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Fence-free one-shot recovery. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Fence-free one-shot recovery.
arXiv: 2608.00501
### Fence-free one-shot recovery.
The negative's subject is the natural discipline (each recoverer reads, arms once, fires once) run twice in parallel without a... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:23"
] | 369 | en | 0.45 | 2026-09-21T09:48:25.066399Z |
arxiv-2608.00501-24 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › What ordering changes. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › What ordering changes.
arXiv: 2608.00501
### What ordering changes.
Suppose each recoverer must prepare its batch, repair the store, and then send, with no step skipped. The model permits only one store repai... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:24"
] | 147 | en | 0.4 | 2026-09-21T09:48:25.066402Z |
arxiv-2608.00501-25 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › The claim fence. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › The claim fence.
arXiv: 2608.00501
### The claim fence.
An atomic *claim* by generation $g$ requires $\varphi\leq g$ , so
claims never lower the fence. The claim sets the fence to $g$ and arms
the batch compu... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:25"
] | 457 | en | 0.475 | 2026-09-21T09:48:25.066404Z |
arxiv-2608.00501-26 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Retention-relative exactness. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Retention-relative exactness.
arXiv: 2608.00501
### Retention-relative exactness.
The theorem's conclusion is graded against what retention kept, and
the difference matters. The machine's retention step has n... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:26"
] | 273 | en | 0.45 | 2026-09-21T09:48:25.066407Z |
arxiv-2608.00501-28 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Interface between the two fences. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 7. Concurrent Recoverers and the Claim Fence › Interface between the two fences.
arXiv: 2608.00501
### Interface between the two fences.
This paper now has two fences, and no theorem merges them. The arrival
fence guards a *wire*: one recoverer, arrivals ... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:28"
] | 272 | en | 0.45 | 2026-09-21T09:48:25.066413Z |
arxiv-2608.00501-29 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 8. Evidence Lifetime | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 8. Evidence Lifetime
arXiv: 2608.00501
## 8. Evidence Lifetime
Systems often combine at-least-once delivery with deduplication (Treat, 2015; Kleppmann, 2017). Theorem 1 gives a coverage identity for an ideal sink view with permanent memory. The remaining ... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:29"
] | 1,139 | en | 0.65 | 2026-09-21T09:48:25.066416Z |
arxiv-2608.00501-30 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 9. Store-Tier Characterization of Log-Derived State | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 9. Store-Tier Characterization of Log-Derived State
arXiv: 2608.00501
## 9. Store-Tier Characterization of Log-Derived State
CDC and the transactional outbox are the standard remedy for dual writes, and they are widely deployed: Debezium alone ships log c... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:30"
] | 1,067 | en | 0.475 | 2026-09-21T09:48:25.066419Z |
arxiv-2608.00501-31 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 10. The Mechanization | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 10. The Mechanization
arXiv: 2608.00501
## 10. The Mechanization
This section explains how the recovery models relate and what the Isabelle proofs establish.
### Relationship between the machines.
Write $\iota$ for the embedding of effect-machine states... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:31"
] | 1,226 | en | 0.675 | 2026-09-21T09:48:25.066422Z |
arxiv-2608.00501-32 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 11. Related Work | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 11. Related Work
arXiv: 2608.00501
## 11. Related Work
### Operational guidance.
The operational rule has been public for a decade, in places peer-reviewed: Kleppmann's dual-writes argument and its book form (Kleppmann, 2015; Kleppmann, 2017), the log-ce... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:32"
] | 759 | en | 0.55 | 2026-09-21T09:48:25.066425Z |
arxiv-2608.00501-33 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery
arXiv: 2608.00501
## 12. A Practitioner's Guide to Dual Writes and CDC Recovery
This section translates the formal results into an operational guide for
backend software engineers, systems archite... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:33"
] | 118 | en | 0.35 | 2026-09-21T09:48:25.066428Z |
arxiv-2608.00501-34 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › The dual-write dilemma across one process and a CDC relay. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › The dual-write dilemma across one process and a CDC relay.
arXiv: 2608.00501
### The dual-write dilemma across one process and a CDC relay.
Dual writes arise whenever an application must record ... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:34"
] | 491 | en | 0.6 | 2026-09-21T09:48:25.066431Z |
arxiv-2608.00501-35 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Defining the acceptance boundary for API sinks and email. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Defining the acceptance boundary for API sinks and email.
arXiv: 2608.00501
### Defining the acceptance boundary for API sinks and email.
To design recoverable delivery, engineers must draw a cl... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:35"
] | 244 | en | 0.425 | 2026-09-21T09:48:25.066435Z |
arxiv-2608.00501-36 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Why local state cannot prove delivery. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Why local state cannot prove delivery.
arXiv: 2608.00501
### Why local state cannot prove delivery.
When a relay crashes and restarts, its local log and outbox records are intact, and the last d... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:36"
] | 255 | en | 0.425 | 2026-09-21T09:48:25.066438Z |
arxiv-2608.00501-37 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Classifying sinks by recovery capability. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Classifying sinks by recovery capability.
arXiv: 2608.00501
### Classifying sinks by recovery capability.
Recovery depends on two distinct capabilities exposed by the sink,
namely what the sink ... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:37"
] | 702 | en | 0.6 | 2026-09-21T09:48:25.066441Z |
arxiv-2608.00501-38 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Controlling in-flight stragglers and concurrent recoverers. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Controlling in-flight stragglers and concurrent recoverers.
arXiv: 2608.00501
### Controlling in-flight stragglers and concurrent recoverers.
Even with queryable or deduplicating sinks, two dist... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:38"
] | 535 | en | 0.55 | 2026-09-21T09:48:25.066443Z |
arxiv-2608.00501-39 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Retention horizons across deduplication caches and source logs. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Retention horizons across deduplication caches and source logs.
arXiv: 2608.00501
### Retention horizons across deduplication caches and source logs.
Deduplication state cannot be retained indef... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:39"
] | 349 | en | 0.5 | 2026-09-21T09:48:25.066447Z |
arxiv-2608.00501-40 | arxiv | https://arxiv.org/abs/2608.00501 | CC-BY-4.0 | Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Architecture checklist, auditing, and recovery tests. | Title: Machine-Checked Dual-Write Recovery from a Commit Log — 12. A Practitioner's Guide to Dual Writes and CDC Recovery › Architecture checklist, auditing, and recovery tests.
arXiv: 2608.00501
### Architecture checklist, auditing, and recovery tests.
To translate these boundaries into practice, teams can apply the... | papers | [
"paper",
"arxiv_id:2608.00501",
"relevance:4",
"chunk:40"
] | 598 | en | 0.525 | 2026-09-21T09:48:25.066450Z |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.