Fábio Silva

Apache Kafka has become a popular tool for building distributed systems. It s a diversity of use cases that benefit from decoupled N-to-M communication such as publishing IoT data, decoupling and load-balancing microservices, and serve as a central hub for data in a distributed application. Nevertheless, Kafka's security is restricted to encrypted communications and authentication, leaving data unprotected in memory and on the disks. In this work, we design and implement a transparent… Mostrar más

Apache Kafka has become a popular tool for building distributed systems. It s a diversity of use cases that benefit from decoupled N-to-M communication such as publishing IoT data, decoupling and load-balancing microservices, and serve as a central hub for data in a distributed application. Nevertheless, Kafka's security is restricted to encrypted communications and authentication, leaving data unprotected in memory and on the disks. In this work, we design and implement a transparent, drop-in component that provides encryption to incoming and outgoing data in a Kafka cluster. Our component leverages confidential computing techniques not only to ensure data-at-rest encryption, but also to protect data and encryption keys from the operators of the Kafka Cluster. Our evaluation shows that the KafkaProxy can handle message streams with latency overhead of around 10%. Finally, in cases where throughput is impacted, simple replication of the KafkaProxy can mitigate the issue. Mostrar menos

Este minicurso apresenta como ferramentas de computação confidencial podem ser usadas para o desenvolvimento de aplicações que processam dados potencialmente sensíveis de aplicações de Internet das Coisas na nuvem. Para o processamento confidencial, nosso minicurso usará tanto o SDK (Software Development Kit) para SGX da Intel (para novas aplicações e serviços) como uma plataforma de execução, SCONE (ideal para execução de aplicações existentes). Para disseminação de dados de Internet das… Mostrar más

Este minicurso apresenta como ferramentas de computação confidencial podem ser usadas para o desenvolvimento de aplicações que processam dados potencialmente sensíveis de aplicações de Internet das Coisas na nuvem. Para o processamento confidencial, nosso minicurso usará tanto o SDK (Software Development Kit) para SGX da Intel (para novas aplicações e serviços) como uma plataforma de execução, SCONE (ideal para execução de aplicações existentes). Para disseminação de dados de Internet das Coisas, nós utilizaremos Apache Kafka e para a orquestração de aplicações, usaremos Kubernetes. Os conceitos apresentados aqui serão ilustrados através de uma aplicação distribuída de processamento de dados de consumo de energia elétrica.

English: This course presents how confidential computing tools can be used to develop applications that process potentially sensitive data from IoT applications in the cloud. For confidential processing, our course will use both the SDK (Software Development Kit) for Intel's SGX (for new applications and services) and an execution platform, SCONE (ideal for running existing applications). For the dissemination of Internet of Things data, we will use Apache Kafka and for the orchestration of applications, we will use Kubernetes. The concepts presented here will be illustrated through a distributed application for processing electricity consumption data. Mostrar menos

Intel SGX has been the subject of numerous research and development projects. Moreover, this technology has been considered a robust option to secure the data being processed in cloud environments. Despite this, configuring SGX-based applications in complex and dynamic scenarios such as microservice architectures is still a challenge. The process of configuring such applications must guarantee the trustworthiness of the services and must be simple and efficient. We then propose a solution for… Mostrar más

Intel SGX has been the subject of numerous research and development projects. Moreover, this technology has been considered a robust option to secure the data being processed in cloud environments. Despite this, configuring SGX-based applications in complex and dynamic scenarios such as microservice architectures is still a challenge. The process of configuring such applications must guarantee the trustworthiness of the services and must be simple and efficient. We then propose a solution for configuring and provisioning secrets to SGX-based applications made with help of the Intel SGX SDK. We present a simple solution that can be easily validated and hardened. Also, the solution is pluggable and can be extended to fit specific requirements or leverage other tools (e.g., for data persistence). In addition to describing our proposal, we also provide an evaluation that shows low overhead to the initialization and configuration time of SGX microservices deployed on Kubernetes. This work contributes to the state-of-the-art of research on using trusted execution environments in cloud computing. Mostrar menos

Cloud computing elastic resources and fine-grained billing enable s to quickly and cost-effectively execute data analysis tasks. Nevertheless, using the cloud instead of on-premise clusters increase the number of ways that sensitive data could be leaked.
In this tutorial, we used the RHD use case from ATMOSPHERE to let the audience understand the necessary steps to implement cloud-based applications that can process sensitive data even when the cloud providers are not trusted.
The… Mostrar más

Cloud computing elastic resources and fine-grained billing enable s to quickly and cost-effectively execute data analysis tasks. Nevertheless, using the cloud instead of on-premise clusters increase the number of ways that sensitive data could be leaked.
In this tutorial, we used the RHD use case from ATMOSPHERE to let the audience understand the necessary steps to implement cloud-based applications that can process sensitive data even when the cloud providers are not trusted.
The tutorial illustrated the anonymization of videos for posterior processing in a way that the original videos are encrypted not only during storage and transmission but even during the processing in the cloud. This tutorial combined the Radiomics application with the SCONE and Asperathos assets. Mostrar menos

This paper presents PubSubSGX, a content-based publish-subscribe system that exploits trusted execution environments (TEEs), such as Intel SGX, to guarantee confidentiality and integrity of data as well as anonymity and privacy of publishers and subscribers.
We describe the technical details of our Python implementation, as well as the required system introduced to deploy our system in a container-based runtime. Our evaluation results show that our approach is sound, while at the… Mostrar más

This paper presents PubSubSGX, a content-based publish-subscribe system that exploits trusted execution environments (TEEs), such as Intel SGX, to guarantee confidentiality and integrity of data as well as anonymity and privacy of publishers and subscribers.
We describe the technical details of our Python implementation, as well as the required system introduced to deploy our system in a container-based runtime. Our evaluation results show that our approach is sound, while at the same time highlighting the performance and scalability trade-offs. In particular, by ing just-in-time compilation inside of TEEs, Python programs inside of TEEs are in general faster than when executed natively using standard ython. Mostrar menos

In this paper we propose a data dissemination platform that s data security and different privacy levels even when the platform and the data are hosted by untrusted infrastructures. The proposed system aims at enabling an application ecosystem that uses off-the-shelf trusted platforms (in this case, Intel SGX), so that s may allow or disallow third parties to access the live data stream with a specific sensitivity-level. Moreover, this approach does not require s to manage the… Mostrar más

In this paper we propose a data dissemination platform that s data security and different privacy levels even when the platform and the data are hosted by untrusted infrastructures. The proposed system aims at enabling an application ecosystem that uses off-the-shelf trusted platforms (in this case, Intel SGX), so that s may allow or disallow third parties to access the live data stream with a specific sensitivity-level. Moreover, this approach does not require s to manage the encryption keys directly. Our experiments show that such an approach is indeed practical for medium scale systems, where participants disseminate small volumes of data at a time, such as in smart grids and IoT environments. Mostrar menos