Divyakant Agrawal

Fuheng Zhao, Divyakant Agrawal, Amr El Abbadi

Database queries traditionally operate under the closed-world assumption, providing no answers to questions that require information beyond the data stored in the database. Hybrid querying using SQL offers an alternative by integrating relational databases with large language models (LLMs) to answer beyond-database questions. In this paper, we present the first cross-domain benchmark, SWAN, containing 120 beyond-database questions over four real-world databases. To leverage state-of-the-art language models in addressing these complex questions in SWAN, we present two solutions: one based on schema expansion and the other based on user defined functions. We also discuss optimization opportunities and potential future directions. Our evaluation demonstrates that using GPT-4 Turbo with few-shot prompts, one can achieves up to 40.0\% in execution accuracy and 48.2\% in data factuality. These results highlights both the potential and challenges for hybrid querying. We believe that our work will inspire further research in creating more efficient and accurate data systems that seamlessly integrate relational databases and large language models to address beyond-database questions.

Ofir Dvir, Kali Hale, Javin Zipkin et al.

We introduce baseSPIDER and SPIDER, private information retrieval (PIR) schemes that embody two technical advancements. The baseSPIDER protocol operates with a single server and a stateful client that performs pre-processing and stores hints for future queries. In this setting, baseSPIDER introduces a new approach that matches the asymptotically optimal communication complexity of state-of-the-art schemes while improving constant factors--an advantage that is particularly significant for databases with large entries. In addition, baseSPIDER offers a conceptually simpler design relative to prior protocols. SPIDER operates over a default database interface and requires no cooperation from the server at any stage. To our knowledge, SPIDER is the first single-server PIR construction of this design, achieving privacy without specialized APIs, auxiliary server state, or protocol-specific interaction beyond conventional indexed access. SPIDER is built via a simple transformation of baseSPIDER to the default server setting, eliminating deployment barriers and enabling immediate applicability to existing systems. This transformation can be applied more broadly to three recent PIR solutions, adapting them for use in the default-server paradigm and yielding solutions of independent interest. SPIDER compares to the resulting modified solutions by exhibiting a simpler design while incurring higher client computational work.

Rithwik Kerur, Divyakant Agrawal, Michael K. Reiter et al.

Spreading and storing erasure-coded data in distributed systems effectively is challenging in real settings. Practical deployments must contend with unpredictable network latencies, particularly when information dispersal is integrated into consensus protocols, a prominent and latency-sensitive use case. Existing approaches address this challenge through timeout-based dissemination and adaptive communication or storage decisions driven by acknowledgments during dissemination. However, these designs focus almost exclusively on dissemination-time efficiency, complicate recovery with reconstruction procedures that require metadata that can differ per consensus value, and rely on a centralized leader to make storage decisions for all nodes. This paper introduces \textbf{Rafture}, a novel information dispersal algorithm, and its integration in a consensus protocol, that overcomes these limitations. Rafture is the first information dispersal solution to incorporate post-dissemination pruning, allowing systems to adapt storage costs after dissemination completes. It employs a simple, fixed-threshold erasure code while varying distinct fragment assignment along a second dimension. This ensures that reconstruction is always possible from $F+1$ fragments using the same interpolation method and no additional metadata. Rafture further enables nodes to adapt autonomously based on locally observed information, eliminating the need for global coordination. We evaluate Rafture in highly dynamic network settings and show that it simplifies recovery while significantly improving long-term storage consumption under variable network conditions.

Fuheng Zhao, Jiayue Chen, Lawrence Lim et al.

Judging the equivalence between two SQL queries is a fundamental problem with many practical applications in data management and SQL generation (i.e., evaluating the quality of generated SQL queries in text-to-SQL task). While the research community has reasoned about SQL equivalence for decades, it poses considerable difficulties and no complete solutions exist. Recently, Large Language Models (LLMs) have shown strong reasoning capability in conversation, question answering and solving mathematics challenges. In this paper, we study if LLMs can be used to determine the equivalence between SQL queries under two notions of SQL equivalence (semantic equivalence and relaxed equivalence). To assist LLMs in generating high quality responses, we present two prompting techniques: Miniature & Mull and Explain & Compare. The former technique is used to evaluate the semantic equivalence in which it asks LLMs to execute a query on a simple database instance and then explore if a counterexample exists by modifying the database. The latter technique is used to evaluate the relaxed equivalence in which it asks LLMs to explain the queries and then compare if they contain significant logical differences. Our experiments demonstrate using our techniques, LLMs is a promising tool to help data engineers in writing semantically equivalent SQL queries, however challenges still persist, and is a better metric for evaluating SQL generation than the popular execution accuracy.

Fuheng Zhao, Jiayue Chen, Yiming Pan et al.

We present the LLM ORDER BY operator as a logical abstraction and study its physical implementations within a unified evaluation framework. Our experiments show that no single approach is universally optimal, with effectiveness depending on query characteristics and data. We introduce three new designs: an agreement-based batch-size policy, a majority voting mechanism for pairwise sorting, and a two-way external merge sort adapted for LLMs. With extensive experiments, our agreement-based procedure is effective at determining batch size for value-based methods, the majority-voting mechanism consistently strengthens pairwise comparisons on GPT-4o, and external merge sort achieves high accuracy-efficiency trade-offs across datasets and models. We further observe a log-linear scaling between compute cost and ordering quality, offering the first step toward principled cost models for LLM powered data systems.

Mohammad Javad Amiri, Joris Duguépéroux, Tristan Allard et al.

Crowdworking platforms provide the opportunity for diverse workers to execute tasks for different requesters. The popularity of the "gig" economy has given rise to independent platforms that provide competing and complementary services. Workers as well as requesters with specific tasks may need to work for or avail from the services of multiple platforms resulting in the rise of multi-platform crowdworking systems. Recently, there has been increasing interest by governmental, legal and social institutions to enforce regulations, such as minimal and maximal work hours, on crowdworking platforms. Platforms within multi-platform crowdworking systems, therefore, need to collaborate to enforce cross-platform regulations. While collaborating to enforce global regulations requires the transparent sharing of information about tasks and their participants, the privacy of all participants needs to be preserved. In this paper, we propose an overall vision exploring the regulation, privacy, and architecture dimensions for the future of work multi-platform crowdworking environments. We then present SEPAR, a multi-platform crowdworking system that enforces a large sub-space of practical global regulations on a set of distributed independent platforms in a privacy-preserving manner. SEPAR, enforces privacy using lightweight and anonymous tokens, while transparency is achieved using fault-tolerant blockchains shared across multiple platforms. The privacy guarantees of SEPAR against covert adversaries are formalized and thoroughly demonstrated, while the experiments reveal the efficiency of SEPAR in terms of performance and scalability.

Sujaya Maiyya, Danny Hyun Bum Cho, Divyakant Agrawal et al.

Significant amounts of data are currently being stored and managed on third-party servers. It is impractical for many small scale enterprises to own their private datacenters, hence renting third-party servers is a viable solution for such businesses. But the increasing number of malicious attacks, both internal and external, as well as buggy software on third-party servers is causing clients to lose their trust in these external infrastructures. While small enterprises cannot avoid using external infrastructures, they need the right set of protocols to manage their data on untrusted infrastructures. In this paper, we propose TFCommit, a novel atomic commitment protocol that executes transactions on data stored across multiple untrusted servers. To our knowledge, TFCommit is the first atomic commitment protocol to execute transactions in an untrusted environment without using expensive Byzantine replication. Using TFCommit, we propose an auditable data management system, Fides, residing completely on untrustworthy infrastructure. As an auditable system, Fides guarantees the detection of potentially malicious failures occurring on untrusted servers using tamper-resistant logs with the support of cryptographic techniques. The experimental evaluation demonstrates the scalability and the relatively low overhead of our approach that allows executing transactions on untrusted infrastructure.

Victor Zakhary, Mohammad Javad Amiri, Sujaya Maiyya et al.

Permissionless blockchains (e.g., Bitcoin, Ethereum, etc) have shown a wide success in implementing global scale peer-to-peer cryptocurrency systems. In such blockchains, new currency units are generated through the mining process and are used in addition to transaction fees to incentivize miners to maintain the blockchain. Although it is clear how currency units are generated and transacted on, it is unclear how to use the infrastructure of permissionless blockchains to manage other assets than the blockchain's currency units (e.g., cars, houses, etc). In this paper, we propose a global asset management system by unifying permissioned and permissionless blockchains. A governmental permissioned blockchain authenticates the registration of end-user assets through smart contract deployments on a permissionless blockchain. Afterwards, end-users can transact on their assets through smart contract function calls (e.g., sell a car, rent a room in a house, etc). In return, end-users get paid in currency units of the same blockchain or other blockchains through atomic cross-chain transactions and governmental offices receive taxes on these transactions in cryptocurrency units.

Victor Zakhary, Divyakant Agrawal, Amr El Abbadi

The recent adoption of blockchain technologies and open permissionless networks suggest the importance of peer-to-peer atomic cross-chain transaction protocols. Users should be able to atomically exchange tokens and assets without depending on centralized intermediaries such as exchanges. Recent peer-to-peer atomic cross-chain swap protocols use hashlocks and timelocks to ensure that participants comply to the protocol. However, an expired timelock could lead to a violation of the all-or-nothing atomicity property. An honest participant who fails to execute a smart contract on time due to a crash failure or network delays at her site might end up losing her assets. Although a crashed participant is the only participant who ends up worse off, current proposals are unsuitable for atomic cross-chain transactions in asynchronous environments where crash failures and network delays are the norm. In this paper, we present AC3WN, the first decentralized all-or-nothing atomic cross-chain commitment protocol. The redeem and refund events of the smart contracts that exchange assets are modeled as conflicting events. An open permissionless network of witnesses is used to guarantee that conflicting events could never simultaneously occur and either all smart contracts in an atomic cross-chain transaction are redeemed or all of them are refunded.

Naman Goel, Divyakant Agrawal, Sanjay Chawla et al.

We propose a new data-centric synchronization framework for carrying out of machine learning (ML) tasks in a distributed environment. Our framework exploits the iterative nature of ML algorithms and relaxes the application agnostic bulk synchronization parallel (BSP) paradigm that has previously been used for distributed machine learning. Data-centric synchronization complements function-centric synchronization based on using stale updates to increase the throughput of distributed ML computations. Experiments to validate our framework suggest that we can attain substantial improvement over BSP while guaranteeing sequential correctness of ML tasks.