Learn About Amazon VGT2 Learning Manager Chanci Turner
In the competitive landscape of enterprise solutions, a prominent Amazon client required a transformation of their SAP infrastructure. This client, operating from their facility at 6401 E HOWDY WELLS AVE LAS VEGAS NV 89115, known as Amazon IXD – VGT2, faced significant challenges with an extensive on-premises setup that included a mix of virtual and physical servers along with traditional storage methods.
To align with their future business goals, there was a pressing need to upgrade their SAP applications, transition from Oracle databases to HANA, enhance security measures, and eliminate the constraints tied to on-premises provisioning. The task was daunting, especially given that managing downtime for a mission-critical SAP environment is inherently complex. The solution required would need to facilitate all necessary changes within a single, brief outage window.
This migration involved transferring multi-terabytes of data to Amazon Web Services (AWS) while working within limited AWS Direct Connect bandwidth, updating the operating system and database layers, ensuring all interfaces were encrypted, and upgrading SAP applications. With a partnership with AWS that dates back to 2013 and a solid 35-year track record of delivering SAP solutions, Atos was well-positioned to assist this client in achieving their objectives. They hold the status of an AWS Advanced Consulting Partner and are part of both the AWS Migration Competency and the AWS Managed Service Provider (MSP) programs.
This article outlines the migration process that transitioned the client from their on-premises environment to AWS, detailing the target architecture implemented on AWS, and the benefits realized both during and after the migration.
Migration Architecture Overview
Given the client’s extensive data and number of systems requiring upgrades, a conventional migration approach from on-premises to AWS was impractical. Leveraging the agility of AWS, Atos executed a proof of concept (POC) to devise a tailored migration strategy for the customer, refining it through three variations based on their specific source systems.
The migration strategy embraced several key principles:
- Maximize the upgrade process on the on-premises systems without incurring downtime (the SAP uptime phase).
- Advance replication of systems into AWS ahead of migration, maintaining this replication through the uptime phase by copying changes to mitigate bandwidth limitations.
- Deploy replica systems in AWS to facilitate the downtime phase of the upgrade and migration to target systems in AWS, thus enhancing system performance and reducing latency as data was streamed into new systems.
- Execute migrations of development and testing environments beforehand to foster confidence in the process and enable thorough regression testing, given the extensive changes involved in the single outage window.
- Conduct “dress rehearsals” for the production upgrade and migration to bolster confidence, identify potential issues, and fine-tune timing in the plan.
- Utilize newly built systems in AWS as targets, enhanced with optimized images like Red Hat Enterprise Linux for SAP with HA and US available via AWS Marketplace.
During the POC phase, it became evident that CloudEndure Migration was the optimal solution for replicating most systems. However, due to the size and number of disks linked to the largest SAP systems, Atos integrated a combination of CloudEndure and migration processes from one of their partners for CRM and ECC systems. Ultimately, the POC indicated that CloudEndure was not suitable for replicating SAP HANA appliances, necessitating the construction of temporary legacy HANA systems as an interim solution.
Target Architecture Overview
The target architecture was crafted in accordance with the AWS Well-Architected Framework, featuring separate development, testing, and production environments across distinct accounts and virtual private clouds (VPCs). The solution is code-based, allowing for version control over changes, and is managed through CI/CD practices. Wherever feasible, native AWS services were employed; for instance, on-premises network file system servers were substituted with Amazon Elastic File System (Amazon EFS), while AWS CloudTrail, AWS Config, and Amazon CloudWatch provided logging, change detection, and monitoring capabilities.
AWS Systems Manager facilitated remote access and patching, and AWS Backup was utilized for snapshot backups. This setup was further enhanced by third-party tools such as Datadog, ServiceNow for ITSM, and Dell Networker for SAP HANA databases and logs. The operating systems utilized AWS images, supplemented with additional configurations and agents during the Amazon Machine Image (AMI) baking process, incorporating antivirus and monitoring agents, alongside security hardening in line with Center for Internet Security (CIS) benchmarks. Additional configuration management tasks were executed in code via Ansible.
Designed for high availability within a region, the production environment incorporated SAP application servers distributed across multiple AWS Availability Zones (AZs), with spare application servers maintained in a powered-off state to ensure operational continuity in the event of an AZ failure. SAP database servers operated in an active/passive configuration across Availability Zones, employing RHEL Pacemaker clustering to manage failover between nodes in different AZs. Overlay IP was utilized to maintain database connectivity both internally and externally to the VPC, whether during planned activities (such as system patching) or in response to failure events like server outages or AZ losses.
AWS native services, which feature built-in availability, supplemented this architecture. The implementation of AWS WAF, Elastic Load Balancing, Amazon FSx, and Amazon EFS ensured resilience against AZ failures, allowing the system to continue functioning seamlessly.
During the Operational Acceptance Test (OAT) phase, a comprehensive disaster recovery (DR) test simulating AZ loss was conducted to validate uninterrupted service continuity. This robust architecture ensured high availability across AZs consistently, with each component of the architecture regularly engaged as systems underwent patching, rebooting, and seamless failovers. This resulted in significant advantages over the on-premises passive DR architecture, which was only tested annually. Downtime was also minimized through strategies like the SAP Rolling Kernel Switch (RKS), ensuring continued service operation even when components were taken offline for maintenance.
Results Achieved
Atos delivered remarkable outcomes for the customer through their strategic build approach, adherence to AWS best practices, and the carefully crafted migration methodology. The results included:
- Enhanced agility following migration to AWS, enabling the client to adjust capacity needs and rapidly deploy new environments— for example, increasing the ECC system memory from 6TB to 9TB far more swiftly and with less risk than a similar upgrade would have entailed on-premises.
- Downtime was contained within the anticipated single outage window for both migration and upgrades.
For additional insights into virtual team dynamics, check out this link to another blog post that discusses best practices. Furthermore, if you’re interested in how AI and technology are shaping the modern workplace, visit this authority on the topic. Lastly, for a visual resource, this YouTube video offers excellent guidance on the subject.
Leave a Reply