Abstract

Abstract
Modern I/O applications that run on HPC infrastructures are increasingly becoming read and metadata intensive. However, having multiple applications submitting large amounts of metadata operations can easily saturate the shared parallel file system's metadata resources, leading to overall performance degradation and I/O unfairness. We present PADLL, an application and file system agnostic storage middleware that enables QoS control of data and metadata workflows in HPC storage systems. It adopts ideas from Software-Defined Storage, building data plane stages that mediate and rate limit POSIX requests submitted to the shared file system, and a control plane that holistically coordinates how all I/O workflows are handled. We demonstrate its performance and feasibility under multiple QoS policies using synthetic benchmarks, real-world applications, and traces collected from a production file system. Results show that PADLL can enforce complex storage QoS policies over concurrent metadata-aggressive jobs, ensuring fairness and prioritization.

Abstract
We present DIO, a generic tool for observing inefficient and erroneous I/O interactions between applications and in-kernel storage systems that lead to performance, dependability, and correctness issues. DIO facilitates the analysis and enables near real-time visualization of complex I/O patterns for data-intensive applications generating millions of storage requests. This is achieved by non-intrusively intercepting system calls, enriching collected data with relevant context, and providing timely analysis and visualization for traced events. We demonstrate its usefulness by analyzing two production-level applications. Results show that DIO enables diagnosing resource contention in multi-threaded I/O that leads to high tail latency and erroneous file accesses that cause data loss.

Abstract
We present DIO, a generic tool for observing inefficient and erroneous I/O interactions between applications and in-kernel storage backends that lead to performance, dependability, and correctness issues. DIO eases the analysis and enables near real-time visualization of complex I/O patterns for data-intensive applications generating millions of storage requests. This is achieved by non-intrusively intercepting system calls, enriching collected data with relevant context, and providing timely analysis and visualization for traced events. We demonstrate its usefulness by analyzing four production-level applications. Results show that DIO enables diagnosing inefficient I/O patterns that lead to poor application performance, unexpected and redundant I/O calls caused by high-level libraries, resource contention in multithreaded I/O that leads to high tail latency, and erroneous file accesses that cause data loss. Moreover, through a detailed evaluation, we show that, when comparing DIO's inline diagnosis pipeline with a similar state-of-the-art solution, our system captures up to 28x more events while keeping tracing performance overhead between 14% and 51%.

Abstract

Abstract
The adoption of third-party machine learning (ML) cloud services is highly dependent on the security guarantees and the performance penalty they incur on workloads for model training and inference. This paper explores security/performance trade-offs for the distributed Apache Spark framework and its ML library. Concretely, we build upon a key insight: in specific deployment settings, one can reveal carefully chosen non-sensitive operations (e.g. statistical calculations). This allows us to considerably improve the performance of privacy-preserving solutions without exposing the protocol to pervasive ML attacks. In more detail, we propose Soteria, a system for distributed privacy-preserving ML that leverages Trusted Execution Environments (e.g. Intel SGX) to run computations over sensitive information in isolated containers (enclaves). Unlike previous work, where all ML-related computation is performed at trusted enclaves, we introduce a hybrid scheme, combining computation done inside and outside these enclaves. The experimental evaluation validates that our approach reduces the runtime of ML algorithms by up to 41% when compared to previous related work. Our protocol is accompanied by a security proof and a discussion regarding resilience against a wide spectrum of ML attacks.