Mlos

Curino, Carlo; Godwal, Neha; Kroth, Brian; Kuryata, S.V.; Lapinski, Greg; Liu, Siqi; Oks, Slava; Poppe, Olga; Smiechowski, Adam; Thayer, Ed; Weimer, Markus; Zhu, Yiwen

doi:10.1145/3399579.3399927

Cited by 7 publications

(1 citation statement)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For Azure Synapse Spark [26], we developed a simulator to mimic the cluster initialization process and derived the optimal policy for sending requests, reducing its tail latency. As another example, by using ML to predict the throughput and latency of benchmark workloads on VMs with various kernel parameters, developed on MLOS [9], we refined the parameters of the Azure VM that runs Redis workloads.…”

Section: Cloud Infrastructure Layermentioning

confidence: 99%

Towards Building Autonomous Data Services on Azure

Zhu,

Tian,

Cahoon

et al. 2023

Companion of the 2023 International Conference on Management of Data

View full text Add to dashboard Cite

Modern cloud has turned data services into easily accessible commodities. With just a few clicks, users are now able to access a catalog of data processing systems for a wide range of tasks. However, the cloud brings in both complexity and opportunity. While cloud users can quickly start an application by using various data services, it can be difficult to configure and optimize these services to gain the most value from them. For cloud providers, managing

show abstract

Section: Cloud Infrastructure Layermentioning

confidence: 99%

Towards Building Autonomous Data Services on Azure

Zhu,

Tian,

Cahoon

et al. 2023

Companion of the 2023 International Conference on Management of Data

View full text Add to dashboard Cite

show abstract

DICE simulation: a tool for software performance assessment at the design stage

et al. 2022

View full text Add to dashboard Cite

In recent years, we have seen many performance fiascos in the deployment of new systems, such as the US health insurance web. This paper describes the functionality and architecture, as well as success stories, of a tool that helps address these types of issues. The tool allows assessing software designs regarding quality, in particular performance and reliability. Starting from a UML design with quality annotations, the tool applies model-transformation techniques to yield analyzable models. Such models are then leveraged by the tool to compute quality metrics. Finally, quality results, over the design, are presented to the engineer, in terms of the problem domain. Hence, the tool is an asset for the software engineer to evaluate system quality through software designs. While leveraging the Eclipse platform, the tool uses UML and the MARTE, DAM and DICE profiles for the system design and the quality modeling.

show abstract

KEA: Tuning an Exabyte-Scale Data Infrastructure

Zhu

Krishnan

Karanasos

et al. 2021

Proceedings of the 2021 International Conference on Management of Data

Self Cite

View full text Add to dashboard Cite

Microsoft's internal big-data infrastructure is one of the largest in the world-with over 300k machines running billions of tasks from over 0.6M daily jobs. Operating this infrastructure is a costly and complex endeavor, and efficiency is paramount. In fact, for over 15 years, a dedicated engineering team has tuned almost every aspect of this infrastructure, achieving state-of-the-art efficiency (>60% average CPU utilization across all clusters). Despite rich telemetry and strong expertise, faced with evolving hardware/software/workloads this manual tuning approach had reached its limit-we had plateaued.In this paper, we present KEA, a multi-year effort to automate our tuning processes to be fully data/model-driven. KEA leverages a mix of domain knowledge and principled data science to capture the essence of our cluster dynamic behavior in a set of machine learning (ML) models based on collected system data. These models power automated optimization procedures for parameter tuning, and inform our leadership in critical decisions around engineering and capacity management (such as hardware and data center design, software investments, etc.). We combine "observational" tuning (i.e., using models to predict system behavior without direct experimentation) with judicious use of "flighting" (i.e., conservative testing in production). This allows us to support a broad range of applications that we discuss in this paper.KEA continuously tunes our cluster configurations and is on track to save Microsoft tens of millions of dollars per year. At the best of our knowledge, this paper is the first to discuss research challenges and practical learnings that emerge when tuning an exabyte-scale data infrastructure.

show abstract

Mlos

Cited by 7 publications

References 13 publications

Towards Building Autonomous Data Services on Azure

Towards Building Autonomous Data Services on Azure

DICE simulation: a tool for software performance assessment at the design stage

KEA: Tuning an Exabyte-Scale Data Infrastructure

Contact Info

Product

Resources

About