Downloaded 51 times
Uploaded byVedanta Barooah
Build and use a DevOps driven Migration Pipeline
This document discusses building and using a DevOps-driven migration pipeline for workload portability between cloud platforms. It outlines several key challenges with application migration including a lack of visibility into application dependencies and architectures. The document then presents HPE Pointnext's workload portability services which use automation and DevOps practices to discover, map, migrate, and optimize workloads for multiple cloud targets. It describes the engagement process and use of tools to rationalize workloads and implement a DevOps style migration pipeline to deploy applications across platforms in a standardized manner.
Build and use a DevOps driven Migration Pipeline
- 2. Build and usea DevOps-driven migration pipeline Automation for Workload Portability Vedanta Barooah – Senior Enterprise Cloud Architect (WW)
- 3. 3 Cloud Workload DeploymentModels – “One size doesn't fit all” Discovery Effort IaaS PaaS Migration Effort Workforce Expertise Vendor Lock-in Open Standards Usage Migration Automation / Tools Cloud Workload Performance TCO & ROI DevOps CI / CD IaaS Migrations (Virtualization Switch) require less effort and low workforce expertise. Customers achieve more automation and cloud performance by using PaaS / Cloud Native Better TCO / ROI using PaaS/Cloud Native Landing zones for Workloads are based on business priorities and often changes with time, its best to sustain them in a composable format which is agnostic towards the target platform
- 4. Common Application MigrationChallenges –Knowledge of current enterprise IT landscape –No single source of truth for application portfolio –Application qualification for Cloud migrations –Applications are continuously moving targets –Requirement to prove ROI before migration –Application performance goals –Multiple landing technologies and targets –Hybrid cloud orchestration and management –New technology and culture adoption challenges
- 5. HPE Pointnext WorkloadPortability Services – Turnkey solution for Cloud Migrations – Start anywhere in the journey – Cloud platform vendor agnostic – Support multiple landing zones – Automation and DevOps at its core – Unique IP and Partner Tools – Includes enablement services for optimization 5 Workload Discovery Identify Workload Suitability Map to Cloud Platforms and Services Migrate Workloads Optimize Workloads Customer Journey for Workload Migration to Cloud
- 6. Workload Suitability andMapping Process Step Model Approach 6 Cloud Suitability − How suitable is a Workload for Cloud ? Hosting Model − What is the Hosting Model ? − Public − Private − Hybrid − Community Platform Suitability − What is the best platform for the workload ? − Helion OpenStack − AWS − Azure − Others Deployment Process − What is the deployment process? − IaaS − SaaS − PaaS Service Mapping − What services within the cloud platforms are used by the workload? − Infrastructure − API − Web Services − OthersWorkload Candidate for Workload Suitability Workload qualified for platform Workload mapped to cloud services Works for both HPE and Non HPE platforms
- 7. Workload Data CollectionStrategies • Define and solidify scope • Use exiting inventory and data sources • Use existing tooling when feasible • Use new tooling when necessary • Multiple options on tools • Reduce noise, discover and filter only required data • Focus on simplicity and speed 7 Scope & Definition Baseline Data Tool Based Discovery Discovery Output Dataset Data Rationalized for Suitability Suitability Existing ToolsCMDB
- 8. Consider what role“Dependency Mapping” plays in transformation ? • Relevant mostly for IaaS deployment model • Server dependencies used for application isolation • Move groups from application dependency • Capacity, Cost and Compute planning from performance matrices • Platform and consolidation planning from Infrastructure inventory • Application service dependency and infrastructure services 8 MySql database MS SQL database Web Site DocumentRoot Web Server Apache Windows Server Guest OS VMware Virtualization Windows Server Host OS Move Groups Consolidation Planning Deployment Model Platform Capacity Planning Cost Planning Compute Planning Application Isolation Active Directory Authentication Dependency Services
- 9. Discovery of workloadinformation is an involved process 9 Infrastructure Server Assets Network Assets Storage Assets Topology Dependency Performance Services Purpose Availability User base Localization Endpoints Access Criteria Software Stack Applications Types Architecture Dependency Business Req. Compliance / SLA Financials Technical Req. Workload Infrastructure Landscape Service Architecture Application Architecture Manual and Automated Rationalizatio n Suitability • Infrastructure discovery Used to identify workload landscape Application Isolation Directly used for IaaS moves Capacity and Performance • Application Discovery Application Dependency Application Architecture Capacity and Performance Attached services Topology Geographic Localization • Services Application Ecosystem Services and Functions Processes
- 10. Workload Migration toCloud - Modules Compute VMware OpenStack AWS EC2 Azure VM Baremetal & Virtual Container Docker AWS ECS Mesos Azure Container Service PaaS AWS Beanstalk AppEngine Stackato Heroku SaaS Enterprise COTS Databases Oracle MySQL MS SQL Server Data Warehouse (TDB) Storage & CDN Storage Migration Cloud Services Public Cloud - AWS - Google Cloud Platform Modernization Enterprise Application Modernization
- 11. HPE Pointnext WorkloadPortability Services 11 Discovery Suitability & Mapping Migration Enablement Workload Evaluation Groups Cloud Mapped Workload Cloud Suitable Evaluation Groups Migrated Workloads Optimized Workloads HPE Workload Rationalization Engine Migration Pipeline Interviews Discovery Tools CMDB Conversion Bridge Workload Optimization Metering & Analytics Automation Optimal Placement
- 12. DevOps Style MigrationPipeline – Engagement Process Migration process integrated with application development lifecycle 12 AURA Pointnext Workload Rationalization Engine DevOps Driven Migration PipelineWorkload Discovery Optimization Standardized Build Templates Automated Migration Multi Cloud Deployment Manual Migration Hybrid Cloud Workload Optimization Your CMDB, systems, or Inventory Data Any Tool based workload discovery Standardized OS Security & Compliance Application Runtime Integrated with Your Continuous Integration, Testing & Deployment Processes Automation to identify suitable workloads Interviews Discovery Tools CMDB Application Migration Data / DB Migration Optimal Performance Build Automation Deployment Automation and Orchestration Optimization Policies Metering & Analytics
- 13. Landing ZoneWorkload State DevOpsPipeline for Workload Migrations 13 Application Review & Changes OS Runtime Application Configuration Deployment Management Engagement = 30% Consulting and Development + 70% Templates and Toolkit Linux (Ubuntu & RHEL) Windows Version #1: LAMP JAVA NodeJS .Net Ruby Python Workloads + Data Migrations Orchestration Inputs from Workload Suitability and Mapping Workload Migration to Cloud Single set of skillsets, tools and processes to deploy across multiple targets Workload Enablement Docker Manifests (20 % HPE Unique IP + 80 % Open Source)
- 14. Landing ZoneWorkload State DevOpsPipeline for Workload Migrations Application Review & Changes OS Runtime Application Configuration Deployment Management Workload Suitability and Mapping Workload Enablement Data collected from Business Communities Asset and Application Data Dev Ops Ops Dev Dev Ops Ops Application Optimization Dev
- 15. Hands on Lab: Objectives – Code Walkthrough – Application deployment on Virtual Machine – Container Build – Container Deployment – Challenge Build 15
- 16. Hands on Lab:Step by Step – You will perform all the steps in a system hosted as a Azure VM – Access system from your desktop using SSH client – Clone source code from - https://github.com/vedanta/dmp – Step by step notes : https://github.com/vedanta/dmp/blob/master/Exercise.md 16
- 17. Thank you Vedanta Barooah(vedanta@hpe.com) 17
Editor's Notes
- #4 Hybrid IT instead of cloud
- #7 Suitability and mapping services enable customers to gauge workload suitability against targeted cloud platforms; this process allows customers to forecast how feasible it is to host an application in the cloud platform and whether there will there be a benefit to migration. Customers can use multiple business, technical, financial, and functional criteria to rationalize workloads against and identify cloud fit. Suitability for workloads can be performed simultaneously for multiple cloud platforms and industry standards. HPE’s unique automated workload rationalization processes allow customers to focus on more critical business and technical priorities. The process of identifying and classifying workloads suitable to multiple industry standard cloud targets is accelerated using HPE’s workload suitability and mapping engine, which automates the cumbersome process of workload rationalization. Using these tools, customers can rapidly create cloud workload portability blueprints, which can be used as a baseline towards successful cloud migration projects. Suitability analysis also takes into account multiple operating and deployment models, and cost benefit analysis to forecast cloud transformation ROI.
- #12 These are the first class citizens. We can support even more than we show here.
- #13 We use a library of components which follow industry best practices based on years of customer engagements. Jenkins, DevOps processes, Saltstack. We don’t disrupt your app dev evolution as migration occurs.