Thiago Garrett
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Thiago Garrett has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
Elias P. Duarte
6 shared publications- Luis Carlos Erpen De Bona
UFPR
6 shared publications - Roman Vitenberg
Department of Informatics, University of Oslo
6 shared publications Ken Birman
2 shared publicationsLeticia M. Peres
2 shared publicationsLigia E. Setenareski
2 shared publications
Latest Publications
- Open Access
- 29 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
On Replacing Cryptopuzzles with Useful Computation in Blockchain Proof-of-Work Protocols
Proof-of-Work (PoW) blockchains have emerged as a robust and effective consensus mechanism in open environments, leading to widespread deployment with numerous cryptocurrency platforms and substantial investments. However, the commonly deployed PoW implementations are all based on solving cryptographic puzzles. Researchers have been pursuing the compelling idea of replacing cryptopuzzles with useful computing tasks for over a decade, in face of the substantial computational capacity of blockchain networks and the global pursuit of a more sustainable IT infrastructure. In this study, we conduct a comprehensive analysis of the prerequisites for alternative classes of tasks. We provide insight into the effect of introducing “usefulness” and of transitioning to task classes other than cryptopuzzles. Having distilled the prerequisites, we use them to examine proposed designs from existing literature. Finally, we discuss pertinent techniques and present research gaps in the current state-of-the-art.
- 29 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
Open AccessOn Replacing Cryptopuzzles with Useful Computation in Blockchain Proof-of-Work Protocols
Proof-of-Work (PoW) blockchains have emerged as a robust and effective consensus mechanism in open environments, leading to widespread deployment with numerous cryptocurrency platforms and substantial investments. However, the commonly deployed PoW implementations are all based on solving cryptographic puzzles. Researchers have been pursuing the compelling idea of replacing cryptopuzzles with useful computing tasks for over a decade, in face of the substantial computational capacity of blockchain networks and the global pursuit of a more sustainable IT infrastructure. In this study, we conduct a comprehensive analysis of the prerequisites for alternative classes of tasks. We provide insight into the effect of introducing “usefulness” and of transitioning to task classes other than cryptopuzzles. Having distilled the prerequisites, we use them to examine proposed designs from existing literature. Finally, we discuss pertinent techniques and present research gaps in the current state-of-the-art.
- Open Access
- 12 March 2026
- Published by Elsevier in Future Generation Computer Systems
- Vol. 182
Transactional feature analysis and synthetic workload generator for UTXO-based DLTs
- 12 March 2026
- Published by Elsevier in Future Generation Computer Systems
- Vol. 182
Open AccessTransactional feature analysis and synthetic workload generator for UTXO-based DLTs
- Open Access
- 27 January 2026
- Published by Association for Computing Machinery (ACM) in Distributed Ledger Technologies: Research and Practice
- Vol. 5 (2) , 1-32
Deconstruction of Knowledge-Building in DAG-Based DLTs
As Distributed Ledger Technologies (DLTs) mature, the inherent performance and scalability shortcomings of linearly structured blockchain designs become better understood. Protocols based on Directed Acyclic Graphs (DAGs) have been proposed to address such shortcomings. DAG-based protocols differ from traditional DLTs in the way they build and represent knowledge about transactions and relations between them. While traditional DLTs have straightforward homogeneous semantic attached to blocks and links between blocks, the semantic of vertices and edges in DAG-based protocols is nuanced and varied. In this work, we identify almost a dozen of knowledge-building dimensions in DAG-based DLTs, none of which have been studied before. Knowledge-building is important in DAG-based DLTs because of its significant impact on the size of the DAG, the pace at which new transactions are added, the finality of transactions, and so on. We analyze 40 DAG-based DLTs from this perspective, summarize our results in a taxonomy, and identify a number of research gaps.
- 27 January 2026
- Published by Association for Computing Machinery (ACM) in Distributed Ledger Technologies: Research and Practice
- Vol. 5 (2) , 1-32
Open AccessDeconstruction of Knowledge-Building in DAG-Based DLTs
As Distributed Ledger Technologies (DLTs) mature, the inherent performance and scalability shortcomings of linearly structured blockchain designs become better understood. Protocols based on Directed Acyclic Graphs (DAGs) have been proposed to address such shortcomings. DAG-based protocols differ from traditional DLTs in the way they build and represent knowledge about transactions and relations between them. While traditional DLTs have straightforward homogeneous semantic attached to blocks and links between blocks, the semantic of vertices and edges in DAG-based protocols is nuanced and varied. In this work, we identify almost a dozen of knowledge-building dimensions in DAG-based DLTs, none of which have been studied before. Knowledge-building is important in DAG-based DLTs because of its significant impact on the size of the DAG, the pace at which new transactions are added, the finality of transactions, and so on. We analyze 40 DAG-based DLTs from this perspective, summarize our results in a taxonomy, and identify a number of research gaps.
- Open Access
Vortex: Hosting ML Inference and Knowledge Retrieval Services With Tight Latency and Throughput Requirements
There is growing interest in deploying ML inference and knowledge retrieval as services that could support both interactive queries by end users and more demanding request flows that arise from AIs integrated into a end-user applications and deployed as agents. Our central premise is that these latter cases will bring service level latency objectives (SLOs). Existing ML serving platforms use batching to optimize for high throughput, exposing them to unpredictable tail latencies. Vortex enables an SLO-first approach. For identical tasks, Vortex's pipelines achieve significantly lower and more stable latencies than TorchServe and Ray Serve over a wide range of workloads, often enabling a given SLO target at more than twice the request rate. When RDMA is available, the Vortex advantage is even more significant.
Open AccessVortex: Hosting ML Inference and Knowledge Retrieval Services With Tight Latency and Throughput Requirements
There is growing interest in deploying ML inference and knowledge retrieval as services that could support both interactive queries by end users and more demanding request flows that arise from AIs integrated into a end-user applications and deployed as agents. Our central premise is that these latter cases will bring service level latency objectives (SLOs). Existing ML serving platforms use batching to optimize for high throughput, exposing them to unpredictable tail latencies. Vortex enables an SLO-first approach. For identical tasks, Vortex's pipelines achieve significantly lower and more stable latencies than TorchServe and Ray Serve over a wide range of workloads, often enabling a given SLO target at more than twice the request rate. When RDMA is available, the Vortex advantage is even more significant.
- Open Access
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-15
Keep Your Friends Close: Leveraging Affinity Groups to Accelerate AI Inference Workflows
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-15
Open AccessKeep Your Friends Close: Leveraging Affinity Groups to Accelerate AI Inference Workflows
- Open Access
- 24 June 2025
- Published by Elsevier in Computer Science Review
- Vol. 58
Decentralization trends in identity management: From federated to Self-Sovereign Identity Management Systems
- 24 June 2025
- Published by Elsevier in Computer Science Review
- Vol. 58
Open AccessDecentralization trends in identity management: From federated to Self-Sovereign Identity Management Systems
- Open Access
Cascade: A Platform for Delay-Sensitive Edge Intelligence
Interactive intelligent computing applications are increasingly prevalent, creating a need for AI/ML platforms optimized to reduce per-event latency while maintaining high throughput and efficient resource management. Yet many intelligent applications run on AI/ML platforms that optimize for high throughput even at the cost of high tail-latency. Cascade is a new AI/ML hosting platform intended to untangle this puzzle. Innovations include a legacy-friendly storage layer that moves data with minimal copying and a "fast path" that collocates data and computation to maximize responsiveness. Our evaluation shows that Cascade reduces latency by orders of magnitude with no loss of throughput.
Open AccessCascade: A Platform for Delay-Sensitive Edge Intelligence
Interactive intelligent computing applications are increasingly prevalent, creating a need for AI/ML platforms optimized to reduce per-event latency while maintaining high throughput and efficient resource management. Yet many intelligent applications run on AI/ML platforms that optimize for high throughput even at the cost of high tail-latency. Cascade is a new AI/ML hosting platform intended to untangle this puzzle. Innovations include a legacy-friendly storage layer that moves data with minimal copying and a "fast path" that collocates data and computation to maximize responsiveness. Our evaluation shows that Cascade reduces latency by orders of magnitude with no loss of throughput.
- Open Access
- 5 February 2022
- Published by Elsevier in Computer Law & Security Review
- Vol. 44
A survey of Network Neutrality regulations worldwide
- 5 February 2022
- Published by Elsevier in Computer Law & Security Review
- Vol. 44
Open AccessA survey of Network Neutrality regulations worldwide
- Open Access
- 1 December 2021
- Published by Elsevier in Computer Networks
- Vol. 200
A Holistic Approach for Locating Traffic Differentiation in the Internet
- 1 December 2021
- Published by Elsevier in Computer Networks
- Vol. 200
Open AccessA Holistic Approach for Locating Traffic Differentiation in the Internet
- 5 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 20 (3) , 2486-2517
Monitoring Network Neutrality: A Survey on Traffic Differentiation Detection
Network neutrality is becoming increasingly important as the global debate intensifies and governments worldwide implement and withdraw regulations. According to this principle, all Internet traffic must be processed without differentiation, regardless of origin, destination, and/or content. Neutrality supporters claim that traffic differentiation can compromise innovation, fair competition, and freedom of choice. However, detecting that an Internet Service Provider (ISP) is not employing traffic differentiation practices is still a challenge. This paper presents a survey of strategies and tools for detecting traffic differentiation on the Internet. After presenting basic neutrality definitions as well as an overview of the worldwide debate, we describe ways that can be used by an ISP to implement traffic differentiation, and define the problem of differentiation detection. This is followed by a description of multiple existing strategies and tools. These solutions differ mainly on how they execute network measurements, the metrics employed, traffic generation techniques, and statistical methods. We also present a taxonomy for the different types of traffic differentiation and the different types of detection. Finally, we identify open challenges and future research directions.
- 5 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 20 (3) , 2486-2517
Monitoring Network Neutrality: A Survey on Traffic Differentiation Detection
Network neutrality is becoming increasingly important as the global debate intensifies and governments worldwide implement and withdraw regulations. According to this principle, all Internet traffic must be processed without differentiation, regardless of origin, destination, and/or content. Neutrality supporters claim that traffic differentiation can compromise innovation, fair competition, and freedom of choice. However, detecting that an Internet Service Provider (ISP) is not employing traffic differentiation practices is still a challenge. This paper presents a survey of strategies and tools for detecting traffic differentiation on the Internet. After presenting basic neutrality definitions as well as an overview of the worldwide debate, we describe ways that can be used by an ISP to implement traffic differentiation, and define the problem of differentiation detection. This is followed by a description of multiple existing strategies and tools. These solutions differ mainly on how they execute network measurements, the metrics employed, traffic generation techniques, and statistical methods. We also present a taxonomy for the different types of traffic differentiation and the different types of detection. Finally, we identify open challenges and future research directions.