Search

Cloud applications are increasingly shifting from large monolithic services to complex graphs of loosely coupled microservices. Despite their benefits, microservices are prone to cascading performance issues, and can lead to prolonged periods of degraded performance. We present Sage, a machine-learning-driven root cause analysis system for interactive cloud microservices that is both accurate and practical. We show that Sage correctly identifies the root causes of performance issues across a diverse set of microservices and takes action to address them, leading to more predictable, performant, and efficient cloud systems. [ABSTRACT FROM AUTHOR]

The DRAM-Based Reconfigurable Acceleration Fabric (DRAF) uses commodity DRAM technology to implement a bit-level, reconfigurable fabric that improves area density by 10 times and power consumption by more than 3 times over conventional field-programmable gate arrays. Latency overlapping and multicontext support allow DRAF to meet the performance and density requirements of demanding applications in datacenter and mobile environments. [ABSTRACT FROM AUTHOR]

Cloud providers host an increasing number of popular applications, on the premise of resource flexibility and cost efficiency. Most of these systems expose virtualized resources of different types and sizes. As instances share the same physical host to increase utilization, they contend on hardware resources, e.g., last-level cache, making them vulnerable to side-channel attacks from co-scheduled applications. In this work we show that using data mining techniques can help an adversarial user of the cloud determine the nature and characteristics of co-scheduled applications and negatively impact their performance through targeted contention injections. We design Bolt, a simple runtime that extracts the sensitivity of co-scheduled applications to various types of interference and uses this signal to determine the type of these applications by applying a set of data mining techniques. We validate the accuracy of Bolt on a 39-server cluster. Bolt correctly identifies the type and characteristics of 81 percent out of 108 victim applications, and constructs specialized contention signals that degrade their performance. We also use Bolt to find the most commonly-run applications on EC2. We hope that underlining such security vulnerabilities in modern cloud facilities will encourage cloud providers to introduce stronger resource isolation primitives in their systems. [ABSTRACT FROM PUBLISHER]

Large-scale datacenters host tens of thousands of diverse applications each day. However, performance is degraded by interference between colocated workloads and the difficulty of matching applications to one of the many hardware platforms available, violating the quality of service (QoS) guarantees that many cloud workloads require. Thus, the authors present Paragon, an online and scalable datacenter scheduler that is aware of heterogeneity and interference. Paragon is derived from robust analytical methods. Instead of profiling each application in detail, it leverages information the system already has about applications it has previously seen. It uses collaborative filtering techniques to quickly and accurately classify an unknown, incoming workload with respect to heterogeneity and interference by identifying similarities to previously scheduled applications. The classification allows Paragon to greedily schedule applications in a manner that minimizes interference and maximizes server utilization. Paragon scales to tens of thousands of servers with marginal scheduling overheads. The authors evaluated Paragon with many workload scenarios, on both small and large-scale systems, including 1,000 servers on Amazon Elastic Compute Cloud (Amazon EC2). For a 2,500-workload scenario, Paragon preserves performance constraints for 91 percent of applications, while significantly improving utilization. In comparison, a baseline least-loaded scheduler only provides similar guarantees for 3 percent of workloads. The differences are more striking during high load when resource efficiency is more critical. [ABSTRACT FROM AUTHOR]

Large-scale datacenters (DCs) host tens of thousands of diverse applications each day. However, interference between colocated workloads and the difficulty of matching applications to one of the many hardware platforms available can degrade performance, violating the quality of service (QoS) guarantees that many cloud workloads require. While previous work has identified the impact of heterogeneity and interference, existing solutions are computationally intensive, cannot be applied online, and do not scale beyond a few applications. We present Paragon, an online and scalable DC scheduler that is heterogeneity- and interference-aware. Paragon is derived from robust analytical methods, and instead of profiling each application in detail, it leverages information the system already has about applications it has previously seen. It uses collaborative filtering techniques to quickly and accurately classify an unknown incoming workload with respect to heterogeneity and interference in multiple shared resources. It does so by identifying similarities to previously scheduled applications. The classification allows Paragon to greedily schedule applications in a manner that minimizes interference and maximizes server utilization. After the initial application placement, Paragon monitors application behavior and adjusts the scheduling decisions at runtime to avoid performance degradations. Additionally, we design ARQ, a multiclass admission control protocol that constrains application waiting time. ARQ queues applications in separate classes based on the type of resources they need and avoids long queueing delays for easy-to-satisfy workloads in highly-loaded scenarios. Paragon scales to tens of thousands of servers and applications with marginal scheduling overheads in terms of time or state. We evaluate Paragon with a wide range of workload scenarios, on both small and large-scale systems, including 1,000 servers on EC2. For a 2,500-workload scenario, Paragon enforces performance guarantees for 91% of applications, while significantly improving utilization. In comparison, heterogeneity-oblivious, interference-oblivious, and least-loaded schedulers only provide similar guarantees for 14%, 11%, and 3% of workloads. The differences are more striking in oversubscribed scenarios where resource efficiency is more critical. [ABSTRACT FROM AUTHOR]

The hundreds of thousands of servers in modern warehouse-scale systems make performance and efficiency optimizations pressing design challenges. These systems are traditionally considered homogeneous. However, that is not typically the case. Multiple server generations compose a heterogeneous environment, whose performance opportunities have not been fully explored since techniques that account for platform heterogeneity typically do not scale to the tens of thousands of applications hosted in large-scale cloud providers. We present ADSM, a scalable and efficient recommendation system for application-to-server mapping in large-scale datacenters (DCs) that is QoS-aware. ADSM overcomes the drawbacks of previous techniques, by leveraging robust and computationally efficient analytical methods to scale to tens of thousands of applications with minimal overheads. It is also QoS-aware, mapping applications to platforms while enforcing strict QoS guarantees. ADSM is derived from validated analytical models, has low and bounded prediction errors, is simple to implement and scales to thousands of applications without significant changes to the system. Over 390 real DC workloads, ADSM improves performance by 16% on average and up to 2.5x and efficiency by 22% in a DC with 10 different server configurations. [ABSTRACT FROM PUBLISHER]

Suboptimal storage design has significant cost and power impact in large-scale datacenters (DCs). Performance, power and cost-optimized systems require deep understanding of target workloads, and mechanisms to effectively model different storage design choices. Traditional benchmarking is invalid in cloud data-stores, representative storage profiles are hard to obtain, while replaying applications in different storage configurations is impractical both in cost and time. Despite these issues, current workload generators are not able to reproduce key aspects of real application patterns (e.g., spatial/temporal locality, I/O intensity). In this paper, we propose a modeling and generation framework for large-scale storage applications. As part of this framework we use a state diagram-based storage model, extend it to a hierarchical representation, and implement a tool that consistently recreates DC application I/O loads. We present the principal features of the framework that allow accurate modeling and generation of storage workloads, and the validation process performed against ten original DC application traces. Finally, we explore two practical applications of this methodology: SSD caching and defragmentation benefits on enterprise storage. Since knowledge of the workload's spatial and temporal locality is necessary to model these use cases, our framework was instrumental in quantifying their performance benefits. The proposed methodology provides detailed understanding of the storage activity of large-scale applications, and enables a wide spectrum of storage studies, without the requirement to access application code and full application deployment. [ABSTRACT FROM PUBLISHER]