Placeholder image

Scalability in the Cloud

| Anbu Govindasamy | Sathish Thirunethiram | Denny Koovakattu |

Infrastructure Scale

YouTube

Anchor

Spotify

Apple

Episode #221

Introduction

In episode 221 of our SAP on Azure video podcast we talk about a huge customer migration project. 200 TB data of Oracle were migrated to SAP on Azure. How was the migration was done? How was the migration downtime minized? To tell us more about this project we have Anbu Govindasamy, Sathish Thirunethiram and Denny Koovakattu joining us!

Find all the links mentioned here: https://www.saponazurepodcast.de/episode221

Reach out to us for any feedback / questions:

#Microsoft #SAP #Azure #SAPonAzure #Oracle

Summary created by AI

  • Introduction and Participant Backgrounds:
  • Goran introduced the meeting, highlighting the focus on a customer migration project to Azure involving a large Oracle SAP system. Danny, Anbu, and Satish introduced themselves, detailing their roles and experience.
    • Goran’s Introduction: Goran introduced the meeting, mentioning the focus on a customer migration project to Azure involving a large Oracle SAP system. He expressed his excitement about discussing the project with Danny, Anbu, and Satish.
    • Participant Introductions: Danny introduced himself as a database architect with over 25 years of experience in the healthcare industry, focusing on performance optimization. Anbu introduced himself as part of the Azure engineering team with 28 years of experience in SAP. Satish introduced himself as an SAP architect working with strategic partnerships at Microsoft, with 18 years of experience in the SAP world.
  • Customer Migration Project Overview:
  • Goran and Satish discussed the engagement with a customer to move a 200 terabyte Oracle database to Azure. They emphasized the importance of understanding the customer’s journey and addressing their pain points.
    • Customer Engagement: Satish explained how they engaged with the customer, emphasizing the importance of understanding the customer’s journey and addressing their pain points. They focused on building a meaningful relationship rather than just selling a product.
    • Initial Assessment: Satish described the initial assessment where a partner had concluded that the 200 terabyte Oracle database could not be moved to any cloud. Microsoft challenged this by requesting the customer’s data to analyze and provide a solution.
    • Data Analysis: Satish detailed the process of analyzing the customer’s data, including AWR reports and system activity reports, to determine the feasibility of moving the workload to Azure. They compared the data against industry peers to validate their solution.
  • Technical Challenges and Solutions:
  • Satish explained the technical challenges faced during the migration, including Oracle licensing, IO, and throughput requirements. He detailed the architecture and solutions proposed to meet these challenges.
    • Oracle Licensing: Satish discussed the challenges related to Oracle licensing, including the need to support Oracle on OEL and Windows. They proposed using Oracle’s own technology to protect the database layer.
    • IO and Throughput: Satish explained the IO and throughput requirements, highlighting the customer’s need for high IO and throughput capabilities. They proposed using ANF and optimizing Oracle parameters to manage these requirements.
    • Architecture Proposal: Satish described the proposed architecture, including moving from IBM power to AVM-based solutions in Azure and using ANF for storage. They aimed to reduce throughput requirements by increasing cache and using ANF to handle IOPS.
  • Azure VM and Storage Innovations:
  • Anbu described the advancements in Azure VM and storage technology, including the use of Intel’s Cascade Lake and DDR5 memory, as well as Azure Boost for improved IO and throughput.
    • Intel Innovations: Anbu detailed the advancements in Intel technology, including the use of Cascade Lake and DDR5 memory, which provided significant improvements in memory access and overall performance.
    • Azure Boost: Anbu explained Azure Boost, which offloads storage and IO traffic to specialized hardware, freeing up the host and increasing IO and throughput capabilities. This innovation allowed for significant improvements in performance.
    • Performance Improvements: Anbu highlighted the performance improvements achieved with the new VM SKU, including increased IOPS and throughput, and the ability to handle large workloads efficiently.
  • Storage Layout and Configuration:
  • Anbu demonstrated the storage layout and configuration using Azure NetApp Files (ANF) and managed disks, highlighting the importance of distributing storage to avoid hotspots.
    • ANF Configuration: Anbu demonstrated the configuration of Azure NetApp Files (ANF), explaining how it supports high throughput and IOPS requirements. He emphasized the importance of distributing storage to avoid hotspots.
    • Managed Disks: Anbu discussed the use of managed disks for hosting binary archive files and temp files, highlighting the benefits of using temp disks for temporary database operations.
    • Template-Based Deployment: Anbu showcased a template-based approach for deploying storage configurations, which simplifies the process and ensures optimal distribution of storage resources.
  • Testing and Validation:
  • Denny discussed the testing and validation process, emphasizing the importance of setting clear goals, identifying critical business windows, and testing for failure to ensure the solution meets business requirements.
    • Testing Strategy: Denny outlined the testing strategy, which included setting clear goals, identifying critical business windows, and using a combination of synthetic and actual business workloads to validate the solution.
    • Failure Testing: Denny emphasized the importance of testing for failure to understand how each component behaves under stress and to identify potential issues before going live.
    • Performance Monitoring: Denny discussed the need to monitor for contention and performance issues, especially as systems run faster in the cloud. He highlighted the differences observed between MV2 and MV3 VMs during testing.
  • Data Protection Strategy:
  • Anbu and Satish outlined the data protection strategy, including the use of ANF snapshots, Oracle Data Guard, and incremental backups to ensure business continuity and minimize downtime.
    • ANF Snapshots: Anbu explained the use of ANF snapshots for quick backups, which do not impact the primary database and can be completed in seconds.
    • Oracle Data Guard: Anbu described the use of Oracle Data Guard to replicate data to a standby system, allowing for offloaded backups and ensuring minimal impact on the primary workload.
    • Incremental Backups: Anbu and Satish discussed the importance of incremental backups and full backups for long-term data protection, ensuring that the system can be restored quickly in case of failure.
  • Project Success and Future Goals:
  • Anbu and Satish concluded by highlighting the success of the project and the strong partnership with the customer. They expressed their commitment to pushing the boundaries further in future projects.
    • Project Success: Anbu and Satish highlighted the success of the project, emphasizing the strong partnership with the customer and the ability to meet all performance and data protection requirements.
    • Future Goals: Anbu and Satish expressed their commitment to pushing the boundaries further in future projects, aiming to handle even larger databases and more complex migrations.

Links

Video index

Based on and using Hugo Cookbook Theme, Font Awesome Icons and Bulma CSS. | Thanks to Recipes4AJ to get me started. | Contact me via