SUMMARY

Data is currently one of the most important assets of our society, given its significance for a wide range of critical services and sectors (e.g., healthcare, e-governance, finance, natural sciences).

While the hardware and software for processing vast and heterogeneous amounts of data have been advancing rapidly, the same is not valid for storage counterparts. However, storage solutions (e.g., block devices, file systems, object stores) provide the cornerstone for persisting and accessing digital information, meaning that their efficiency is paramount to fully leveraging the advancements in data processing.

Our research aims to build efficient storage solutions tailored to the performance, dependability, and security demands of emerging applications (e.g., AI, analytics, databases) and large-scale infrastructures (e.g., cloud computing, high-performance computing).

OBJECTIVES

Within this domain, our current areas of interest include:

  • Adaptable and extensible storage systems: We research solutions that break from the monolithic design of traditional storage systems, which struggle to accommodate heterogeneous and emerging workloads. By building extendable storage systems that incorporate novel optimizations and features, chosen and tailored at runtime, we aim to improve data management for a broader range of applications and services.
  • Transparent and tailored storage features: We research storage features individually (e.g., load-balancing, caching, deduplication, scheduling) and at scale to ensure better management of storage resources and better quality of service for cloud computing and HPC infrastructures. As a fundamental principle, we design transparent and non-intrusive optimizations for the different applications leveraging our solutions.
  • Dependability and security: We design storage systems that are resilient and adaptable to environments with different levels of instability and failures (e.g., private servers, data centers, IoT). Also, we design efficient solutions that transparently ensure the privacy and security of data (e.g., encryption at rest and in transit, trusted hardware ) when stored in untrusted third-party services and under threats (e.g., ransomware attacks).

SELECTED PUBLICATIONS

  • Taming Metadata-intensive HPC Jobs Through Dynamic, Application-agnostic QoS Control.
    Macedo R, Miranda M, Tanimura Y, Haga J, Ruhela A, Harrell S, Evans T, Pereira J, Paulo J.
    IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing (CCGrid), 2023.
  • PAIO: General, Portable I/O Optimizations With Minor Application Modifications.
    Macedo R, Tanimura Y, Haga J, Chidambaram V, Pereira J, Paulo J..
    USENIX Conference on File and Storage Technologies (FAST), 2022.
  • SafeFS: a modular architecture for secure user-space file systems: one FUSE to rule them all.
    Pontes R, Burihabwa D, Maia F, Paulo J, Schiavoni V, Felber P, Mercier H, Oliveira R.
    ACM International Systems and Storage Conference (SYSTOR) (best student paper award), 2017.
Check the full list of publications here.

RESPONSIBLE FOR THE DOMAIN

João Paulo

Assistant Professor
(Coordinator)