Positions: 10

Bolsa de Introdução à; Investigação

  • Project Center for Responsible AI – refª C628696807-00454142

    Type of position: Bolsa de Introdução à; Investigação
    Type of contract: Research grant
    Duration: 3 months
    Deadline to apply: 2024-03-28
    Description

    The main goal is to build a Proof of Space blockchain that relies on real contents that are maintained in the nodes of Machine Learning, Data

    Analytics and Cloud platforms.

     

    More precisely, the work will integrate previous work in our team on the Proof of Space blockchain protocol, called “Chia Bread Pudding”,

    with an encrypted file system on a mainstream Operating System, such as Linux. In this blockchain protocol, the stake of each user is given by the amount of file contents that he holds in his disks.

    The goal of the integrated system that should result from this grant is to allow “Chia Bread Pudding” to use the user’s contents that are

    already maintained in an encrypted file system. The implemented solution should also try to ensure that this integration does not incur relevant overheads on the normal operation of the file system.

     

    The work plan will include:

     

    Study available file systems and choose one that meets the requirements of this project.

     

    Build the integrated system, which should include measures to minimise the overheads.

     

    Experimentally evaluate the resulting system, using realistic workloads from relevant benchmarks from the ML/data analytics/cloud domains.

     

    The expected outcome is a design and implementation of the above-mentioned integrated systems, together with a comprehensive experimental evaluation, to be presented in at least one scientific paper.


    Contact email: rh@inesc-id.pt

Bolsa de Investigação

  • MS3 – Refª PTDC/CCI-INF/6762/2020-BI|2024/509-BI|2024/510

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 3 months
    Deadline to apply: 2024-04-05
    Description

    BI|2024/509

    One of the scholarship holders will focus on the development of a benchmark for micro-services that has the characteristic of being realistic (leverage existing large-scale traces), complete (incorporate information such as the type of operation and the object ids that allow for identifying the degree of sharing and concurrency), and easy to deploy, so that it can be made available as open source.

     

    BI|2024/510

    The other scholarship holder will try to incorporate previous concepts developed by the research team, namely specific micro-services that offer resilience against unfavorable node or network conditions, into real-world production systems.


    Contact email: rh@inesc-id.pt
  • CRAI - refª C628696807-00454142 - BI|2024/496

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 12 months
    Deadline to apply: 2024-03-29
    Description

    State-of-the-art AI models are unsustainably hunger for data and computational resources: the recent GPT-3 model developed by OpenAI has 175 billion parameters, was trained on 500 billion words, and took 3.14E23 flops to train (on a domestic GPU, it would cost $4.6M and it would require 355 years to train such a model). This has a daunting carbon footprint and poses serious environmental concerns. Recent literature has shown that the training of a state of the art ML model for NLP can emit 5x more CO2 than a car during its entire lifespan.

    The assignee of this research grant will investigate and develop new scientific methods and technologies for making AI systems more energy and data efficient an important goal for Data4Green and the European Green Deal.

    More in detail, the focus of this grant will be on investigating the adoption of emerging hardware mechanisms (In-memory processing devices, Non-Volatile RAM, Transactional Memory) to increase efficiency of storage and intra-node synchronization.


    Contact email: rh@inesc-id.pt
  • AINUR – Refª PTDC/CCI-COM/4485/2021 - BI|2024/515 and 516

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-28
    Description

    BI|2024/515

    The goal of this work is to redesign SPHT (the state-of-the-art solution for persistent memory transactions relying on the previous version of Intel TSX, proposed by our team at INESC-ID) to take into account two technological trends: (1) the opportunity of the new transactional instructions that are now available in the latest generations of Intel CPUs (Sapphire Rapids and Emerald Rapids); (2) the restriction that Intel Optane Persistent Memory has been discontinued, so that the new systems must either resort to fast PCIe-based SSD storage or, in the future, to CXL-based persistent memories.

    The proposed solution is expected to ensure crash consistency guarantees, will also be tested using the tools developed by the Ainur project.

    The work plan will include:

    a) Design a new algorithm for persistent hardware transactions for the above-mentioned trends, borrowing ideas from SPHT.

    b) Implement and evaluate the new system, using realistic workloads from relevant benchmarks and focusing on the two possible scenarios for the persistent memory (SSD and CXL).

    c) Test the proposed solution using tools from the Ainur project.

    The expected outcome is a design and implementation of the above-mentioned system, together with a comprehensive experimental evaluation, to be presented in at least one scientific paper.

     

    BI|2024/516

    The goal of this work is to study the synchronization mechanisms that eBPF provides to programs that need to concurrently access data structures maintained in shared memory, and to propose improved mechanisms that exploit more sophisticated techniques from literature – including wait-free data structures and hardware transactional memory.

    Complementarily, we will consider eBPF programs that store the shared data structures durably on persistent memory.

    This work will mainly focus on highly concurrent eBPF workloads, for which synchronization is a major challenge.

    The work plan will include:

     

    a) Experimentally study the performance and progress trade-offs of the existing synchronisation support of eBPF, which is based on spin locks.

     

    b) Design, implement and evaluate at least one alternative synchronization mechanism for eBPF. The evaluation will use realistic high-contention workloads from relevant eBPF benchmarks and publicly available eBPF programs. Furthemore, the evaluation should also consider scenarios where the shared data structures are maintained persistently in persistent memory (such as Intel Optane DC PM).

     

    The expected outcome is a design and implementation of the above-mentioned system, together with a comprehensive experimental evaluation, to be presented in at least one scientific paper.


    Contact email: rh@inesc-id.pt
  • Project DikesFPro-Refª PTDC/ECI-EGC/7739/2020-BI|2024/513

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-27
    Description

    The DikesFPro project is currently focusing on citizen science and the assimilation of contributions from citizen volunteers (e.g., students, researchers, or environmental activists). In this context, the grantee will focus on advancing the RiverCore Portal, an established platform for flood simulation processes, to enable citizens to contribute field data during environmental events (e.g., such as floods, drought, contamination, and water quality). This entails the entire process of design, implementation, and testing user experience with various stakeholders.


    Contact email: rh@inesc-id.pt
  • Project DikesFPro - Refª PTDC/ECI-EGC/7739/2020-BI|2024/514

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-27
    Description

    The DikesFPro project is currently enhancing the RiverCure Portal, an established platform designed to facilitate flood simulation processes. As part of this initiative, the grantee will focus on advancing this platform to enable its users to create and manage multiple versions of geographic contexts (which currently is limited to the current version only) to improve the analysis and decision-support processes in what concerns safety and risk management of the water basins contexts during environmental events (e.g., such as floods, drought, contamination, and water quality).


    Contact email: rh@inesc-id.pt
  • SmartRetail – Refª C6632206063-00466847 - BI|2024/511

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-26
    Description

    This project aims to enhance the security and privacy of smart-retail environments through cutting-edge technology. Our dual objectives begin with a comprehensive examination of the network architecture employed by contemporary smart-retail stores, identifying potential vulnerabilities and inefficiencies. Following this, we will design and implement a network security solution that can guarantee the protection of confidentiality, integrity, and availability of in-house and cloud-based systems that support the smart-retail data processing pipeline. This solution will explore the integration of adaptable switches and the implementation of specialized data plane software utilizing the P4 programming language, known for its ability to program packet-forwarding planes.


    Contact email: rh@inesc-id.pt
  • Project SmartRetail – Refª C6632206063-00466847-BI|2024/512

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-26
    Description

    The student will become an integral part of the SmartRetail project with a focus on the development of secure web applications essential for smart retail store operations. The role will primarily involve leveraging Explode.js for TypeScript applications. Originally, Explode.js was crafted to pinpoint security vulnerabilities within JavaScript applications. This initiative aims to adapt and extend its framework for identifying security flaws in TypeScript environments, encompassing both client and server-side code. The project is structured around three main goals: (1) perform an exhaustive analysis of existing methods for detecting TypeScript vulnerabilities, (2) engineer and prototype a tool for TypeScript vulnerability detection based on Explode.js's capabilities, and (3) rigorously test and assess the efficacy of the prototype tool. Expected outcomes include a functioning software prototype and a comprehensive technical report detailing the findings and development process.


    Contact email: rh@inesc-id.pt
  • ATE – refª C644914747-00000023- BI|2024/507

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 6 months
    Deadline to apply: 2024-03-20
    Description
    State of the art of Energy Hubs. Model Energy Hubs considering different energy resources. Implement energy management models. Propose a protection system for energy hubs. Test and validation. Assist the research team in the production of reports and research papers


    Contact email: rh@inesc-id.pt
  • SALAD-Nets – refª 2022.15622.CMU-BI|2024/508

    Type of position: Bolsa de Investigação
    Type of contract: Research grant
    Duration: 9 months
    Deadline to apply: 2024-03-20
    Description

    The SALAD-Nets project aims to automatically synthesize networks in a large-scale cloud/data-center environment, using programmable network equipment. We will build on our prior work [1, 2] to analyze large virtual network topologies, connecting servers and VMs through a variety of software network functions (NFs), and the physical network fabric topology that the virtual network will be deployed onto. We will process this to generate all of the network programs and controllers required to implement and deploy the virtual network using the physical fabric.

     As a part of this fellowship, you will build a proof-of-concept prototype synthesis engine to (i) analyze and represent the virtual

    network in a scalable manner, (ii) to optimally map virtual network functionality onto the physical network equipment, and (iii) to generate the final network programs and controllers that realize the virtual network on the physical fabric.


    Contact email: rh@inesc-id.pt