The document summarizes a conference called Velocity 2013 Conference. It provides details about:
- The purpose of attending the conference was to learn about reference architectures for build/deployment systems, monitoring, and mobile deployment.
- The conference covered topics like mobile performance, operations, and web performance. It included technical sessions on solving performance and operational issues in large-scale environments.
- Some of the sessions discussed tools for packaging, log collection, monitoring metrics, testing methods, lessons learned from projects, and scaling load balancing infrastructure to over a billion users.
I Know It Was MEAN, But I Cut the Cord to LAMP AnywayAll Things Open
All Things Open 2014 - Day 2
Thursday, October 23rd, 2014
Brian Hyder
Co-Founder & CTO of PencilBlue, LLC
Back Dev
I Know It Was MEAN, But I Cut the Cord to LAMP Anyway
JavaOne 2015 Devops and the Darkside CON6447Steve Poole
So you get DevOps. You like the idea and think it’s important. The trouble is that others in your team don’t. This session will help you understand how to convince your team of the benefits of DevOps. Packed with facts and figures, the presentation works through the common challenges Java teams face when moving to a DevOps model and outlines how to address them. It also shows you how to balance evangelism against pragmatism when championing DevOps in your organization. You’ll learn how others have made the transition to DevOps and understand what mistakes to avoid when doing so. Whether you need to know how to be a DevOps evangelist or simply want to understand why DevOps is important, this session is for you.
The Quest for Continuous Delivery at PluralsightMike Clement
Continuous integration, continuous delivery, continuous deployment. These may seem like unreachable goals in your current situation. Maybe you’re doing weekly or even monthly coordinated releases. At Pluralsight we deploy multiple times a week, sometimes multiple times a day. We’re not perfect and we’re constantly working to improve our system, but I’ll share an experience report of where we currently are and where we’re going.
DataOps for the Modern Data Warehouse on Microsoft Azure @ NDCOslo 2020 - Lac...Lace Lofranco
Talk Description:
The Modern Data Warehouse architecture is a response to the emergence of Big Data, Machine Learning and Advanced Analytics. DevOps is a key aspect of successfully operationalising a multi-source Modern Data Warehouse.
While there are many examples of how to build CI/CD pipelines for traditional applications, applying these concepts to Big Data Analytical Pipelines is a relatively new and emerging area. In this demo heavy session, we will see how to apply DevOps principles to an end-to-end Data Pipeline built on the Microsoft Azure Data Platform with technologies such as Data Factory, Databricks, Data Lake Gen2, Azure Synapse, and AzureDevOps.
Resources: https://aka.ms/mdw-dataops
Overseeing Ship's Surveys and Surveyors Globally Using IoT and Docker by Jay ...Docker, Inc.
Fugro is a multinational enterprise that collects and provides highly specialized interpretation of geological data for a number of industries, at land and at sea. The company recently launched OARS (Office Assisted Remote Services), an innovation which uses advanced technology to reduce, and potentially eliminate, the need for surveyors onboard sea-going vessels, optimizing project crewing, safety and efficiency. By keeping skilled staff onshore and using an Internet of Things platform model, Fugro’s OARS project provides faster interpretation of data and decisions, better access to information across regions Hear how Fugro and consulting partner Flux7 created a solution with Docker and Amazon Web Services at its center that provides a high degree of uptime, ensures data is secure and enables portability so that environments that can be quickly replicated in new global regions on demand. Learn how Docker is being used as a key component in Fugro’s continuous delivery cycle and see how Docker is also used to create redundancy that ensures high uptime for Fugro’s 24X7 requirements.
I am an instructor of the MLOps workshop for some anonymous startup incubation program where the objectives are (1) to orchestrate and deploy updates to the application and the deep learning model in a unified way. (2) To design a DevOps pipeline to coordinate retrieving the latest best model from the model registry, packaging the web application, deploying the web application and inferencing web service.
I Know It Was MEAN, But I Cut the Cord to LAMP AnywayAll Things Open
All Things Open 2014 - Day 2
Thursday, October 23rd, 2014
Brian Hyder
Co-Founder & CTO of PencilBlue, LLC
Back Dev
I Know It Was MEAN, But I Cut the Cord to LAMP Anyway
JavaOne 2015 Devops and the Darkside CON6447Steve Poole
So you get DevOps. You like the idea and think it’s important. The trouble is that others in your team don’t. This session will help you understand how to convince your team of the benefits of DevOps. Packed with facts and figures, the presentation works through the common challenges Java teams face when moving to a DevOps model and outlines how to address them. It also shows you how to balance evangelism against pragmatism when championing DevOps in your organization. You’ll learn how others have made the transition to DevOps and understand what mistakes to avoid when doing so. Whether you need to know how to be a DevOps evangelist or simply want to understand why DevOps is important, this session is for you.
The Quest for Continuous Delivery at PluralsightMike Clement
Continuous integration, continuous delivery, continuous deployment. These may seem like unreachable goals in your current situation. Maybe you’re doing weekly or even monthly coordinated releases. At Pluralsight we deploy multiple times a week, sometimes multiple times a day. We’re not perfect and we’re constantly working to improve our system, but I’ll share an experience report of where we currently are and where we’re going.
DataOps for the Modern Data Warehouse on Microsoft Azure @ NDCOslo 2020 - Lac...Lace Lofranco
Talk Description:
The Modern Data Warehouse architecture is a response to the emergence of Big Data, Machine Learning and Advanced Analytics. DevOps is a key aspect of successfully operationalising a multi-source Modern Data Warehouse.
While there are many examples of how to build CI/CD pipelines for traditional applications, applying these concepts to Big Data Analytical Pipelines is a relatively new and emerging area. In this demo heavy session, we will see how to apply DevOps principles to an end-to-end Data Pipeline built on the Microsoft Azure Data Platform with technologies such as Data Factory, Databricks, Data Lake Gen2, Azure Synapse, and AzureDevOps.
Resources: https://aka.ms/mdw-dataops
Overseeing Ship's Surveys and Surveyors Globally Using IoT and Docker by Jay ...Docker, Inc.
Fugro is a multinational enterprise that collects and provides highly specialized interpretation of geological data for a number of industries, at land and at sea. The company recently launched OARS (Office Assisted Remote Services), an innovation which uses advanced technology to reduce, and potentially eliminate, the need for surveyors onboard sea-going vessels, optimizing project crewing, safety and efficiency. By keeping skilled staff onshore and using an Internet of Things platform model, Fugro’s OARS project provides faster interpretation of data and decisions, better access to information across regions Hear how Fugro and consulting partner Flux7 created a solution with Docker and Amazon Web Services at its center that provides a high degree of uptime, ensures data is secure and enables portability so that environments that can be quickly replicated in new global regions on demand. Learn how Docker is being used as a key component in Fugro’s continuous delivery cycle and see how Docker is also used to create redundancy that ensures high uptime for Fugro’s 24X7 requirements.
I am an instructor of the MLOps workshop for some anonymous startup incubation program where the objectives are (1) to orchestrate and deploy updates to the application and the deep learning model in a unified way. (2) To design a DevOps pipeline to coordinate retrieving the latest best model from the model registry, packaging the web application, deploying the web application and inferencing web service.
Incremental development is easy when we are talking about functionality. Story splitting has become quite popular as a technique lately.
But what about those cases when you need to do an architectural refactoring? Could incremental development be applied?
(Talk delivered during I T.A.K.E. Unconference 2015)
En DevOps är en roll i teamet som överbryggar utveckling och drift och som hjälper till att säkerställa leverenspresission och kvalitet, bland annat genom att se till att bra lösningar för applikationsövervakning, felrapportering, paketering och deployment finns på plats.
1. Velocity 2013 Conference
개요
작성자: 황상철
소속: SK Planet > CTO > SQE 팀
작성일: 2013년7월15일
참가 목적
빌드 & 배포 시스템에 대한 레퍼런스 및 관련 오픈소스 파악
모니터링에 대한 레퍼런스 아키텍처 발굴
모바일 배포에 대한 레퍼런스 및 아이디어
운영에 대한 트렌드 파악
성능에 대한 최신 정보 습득
Velocity 2013 컨퍼런스란?
기간: 2013년 6월18일~20일
장소: 산타클라라, 캘리포니아
주최: 오렐리
사이트: http://velocityconf.com/velocity2013/
참가자 규모: 약 1800명
구성: 튜토리얼, 세션, 부스 전시
컨퍼런스 요약
'Building a Faster and Stronger Web'이라는 캐치프레이즈로 열린 본 컨퍼런스는 크게 3가지 주제를 다루고 있다.
모바일 성능: 네이티브웹,웹앱, 반응형 디자인, 모바일 앱 성능 측정 도구, 프레임워크 등
운영: 모니터링, 지표, 설정관리, DevOps
웹 성능: 최신 도구,기술, 성능 개선을 위한 경험 공유
대용량 환경에서 발생할 수 있는 성능이나 운영상의 문제를 어떻게 해결했는가에 대한 기술적인 내용들
관련 오픈소스, 상용도구, 자체 개발한 시스템에 대한 소개
성능 개선이 비즈니스에 어떻게 도움이 되었는가에 대한 내용
Operations Level Up
http://velocityconf.com/velocity2013/public/schedule/detail/28203
« ‹ › »
1 /123
Mwalls velocity levelup from Mandi Walls
Topics
Tools for building operating system packages for your own deployments
Log collection and aggregation
Metrics, monitoring, and statistics
Working with software development tools, version control, code review
Testing methods, configuration management tools
Lessons and Learned
Git
2. Ops Identity Crisis: 저는 시스템 관리자라 코드를 못 짜는데요.
Easy to do Right
Good Comment, Prc Commit Hook
서버 설정등에 대한 테스트를 자동화하라.
Bring the Noise: Making Effective Use of a Quarter Million Metrics
http://velocityconf.com/velocity2013/public/schedule/detail/28177
발표자료: https://speakerdeck.com/astanway/bring-the-noise-continuously-deploying-under-a-hailstorm-of-metrics
Topics
Metric Overload: The situation which gave rise to the development of these tools and how we approached the problem
Skyline: The architecture and algorithms we use for realtime anomaly detection on a massive scale
Oculus: The architecture and algorithms we use to compute similarity and correlation across all of our metric data
Lessons and Learned
Etsy: http://www.etsy.com/
1.5 billion pv, 950 thousands users
Continuous Delivery
30 deploy a day? Is that safe?
실수를 기다리지 말고 문제를 미리 찾자
Can't fix what you don't measure
리얼타임으로 에러 로그를 분석해서 (잠재)에러를 찾는다.
Kale: Skyline + Oculus
Skyline
A real time anomaly detection system
https://github.com/etsy/skyline
Oculus:
anomaly correlation component
https://github.com/etsy/oculus
Realtime의 의미
Graphite, Ganglia, Redis
Euclidian Distance
Speeding Up Your Mobile HTML5 Experience
http://velocityconf.com/velocity2013/public/schedule/detail/28314
발표자료: https://speakerdeck.com/firtman/speeding-up-your-mobile-html5-experience
Topics
CSS Sprites vs inline images
Animate 3D transforms for hardware acceleration
Application Cache
Different viewport definitions
JavaScript frameworks usage
Web Storage vs. SQL Storage vs. IDB
Large DOM vs iframes vs object pool
SVG vs high resolution canvas
Mouse vs. Touch/Pointer events
Animation Timing API
Images vs CSS3 effects and gradients
Using HTML5 abilities to improve performance perception and responsiveness
Managing PostgreSQL with Ansible in EC2
http://velocityconf.com/velocity2013/public/schedule/detail/28115
Topics
Ansible: https://github.com/ansible/ansible
Ansible is a radically simple configuration-management, deployment, task-execution, and multinode orchestration framework.
Avoiding Performance Regression at Twitter
http://velocityconf.com/velocity2013/public/schedule/detail/28455
3. « ‹ › »
1 /52
Velocity Conf 2013 Workshop: Avoiding Web Performance Regression from Marcel Duran
Topics
YSlow
WebPageTest
http://www.webpagetest.org/
https://github.com/WPO-Foundation/webpagetest
Meet the Obama Campaign's $250 Million Fundraising Platform
http://velocityconf.com/velocity2013/public/schedule/detail/28316
Video
http://kylerush.net/blog/meet-the-obama-campaigns-250-million-fundraising-platform/
Lesson and Learned
성능지표 vs 비즈니스
Is The Web Getting Faster?
http://velocityconf.com/velocity2013/public/schedule/detail/30013
Video
Lesson and Learned
과거에 비해 현재 웹 기술이 얼마나 빨라졌는가
자바스크립트 20배
API Design Principles for Accelerated Development
http://velocityconf.com/velocity2013/public/schedule/detail/31356
« ‹ › »
1 /27
API Design Principles for Accelerated Development from Jonathan LeBlanc
Optimizing the Critical Rendering Path for Instant Mobile Websites
4. http://velocityconf.com/velocity2013/public/schedule/detail/29616
발표자료: https://docs.google.com/presentation/d/1IRHyU7_crIiCjl0Gvue0WY3eY_eYvFQvSfwQouW9368/present#slide=id.p19
Lesson and Learned
네트워크 망 부터 확인하기 시작
TCP 연결을 고려
Critical Path를 찾아서 최적화
http://modpagespeed.com/
http://ngxpagespeed.com/ngx_pagespeed_example/
Increasing Speed To Market In Mobile Development Through Continuous
Integration
http://velocityconf.com/velocity2013/public/schedule/detail/31073
Lesson and Learned
모바일 배포 및 테스트 자동화
실제 장비에 대한 배포
The Secret of Safe, Continuous Deployment: Perceptual Diffs
http://velocityconf.com/velocity2013/public/schedule/detail/28452
발표자료:
http://cdn.oreillystatic.com/en/assets/1/event/94/The%20Secret%20of%20Safe,%20Continuous%20Deployment_%20Perceptual%20Diffs%20Presentation%201.pdf
Lesson and Learned
How perceptual diffs make it safe for web-based software
Why it’s not safe in general to deploy on commit
The stumbling blocks we’ve hit and mistakes we’ve made
Lesson and Learned
Perceptual diffs are the best thing!
과거 버전 스크린 샷과 새로운 버전의 대한 스크린 샷을 비교하는 방법
색상은 빨간색으로 변환해서 비교한다.
관련 오픈소스: https://github.com/bslatkin/dpxdt
The CSS and GPU Cheatsheet
http://velocityconf.com/velocity2013/public/schedule/detail/31377
발표자료: http://cdn.oreillystatic.com/en/assets/1/event/94/The%20CSS%20and%20GPU%20Cheatsheet%20Presentation.pdf
Lesson and Learned
GPU + Layers = faster rendering
Too many layers = seriously bad time
Be aware of your CSS tags impact on CPU & GPU perf
Use tooling to show off how many layers you have
Building A Billion User Load Balancer
http://velocityconf.com/velocity2013/public/schedule/detail/28410
Topics
how facebook scales their load balancing infrastructure to support more than a billion users?
Our Python decision engine is named Cartographer
Lesson and Learned
Context
일주일중에는 월요일이 피크
하루중 11시와 3시가 피크
타입존에 따라 하루중 피크타임이 달라진다.
Demo를 서울에서 오레곤으로 휴가를 오기위해 접속
Scale Out
LB 레이어
TCP 세션 레이어
도쿄에 프록시를 두고 미국서버에 접속할 필요없이 처리해서 성능을 올렸다
Sonar를 활용해서 사용자 프록시에 히트하는지를 검사
Building A Performance Lab For Mobile Apps In A Day
http://velocityconf.com/velocity2013/public/schedule/detail/31041
발표자료: http://www.slideshare.net/ashraymathur/velocity-building-a-performance-lab-for-mobile-apps-in-a-day-final
Topics
How to define and set baselines
Performance metrics and instrumentations
How to keep the performance optimized from code check-in to rollout
5. Internal Monitoring to identify problems and avoid regressions
Real User Monitoring to gauge customer experience
App Crashes – How to detect and handle app crashes & memory leaks
Tools – Some very important tools that need to be in your toolbox
Perceived Performance – Importance of adding performance in app design
Lesson and Learned
Metrics: Page Load Times, Network Metrices, Nav Timing Metrics, Server/Client Metrics
3 Esstntial Steps
Mobile Automation
App Instrumentation
Mobile Infrastructure
CI
Launch Simulator and run tests
Selenium for native iOS apps: https://github.com/moredip/Frank
Stress/longevity testing
Instrumentaion: https://github.com/lognormal/boomerang/
Monitoring gc log, heap dumps(by MAT)
DevOps: It’s Not Just For WebOps And We Have The Metrics To Prove It
http://velocityconf.com/velocity2013/public/schedule/detail/28446