Azure to AWS Migration and Database Optimization

IRON is a leader in Global OEM Integration Services for server appliances and cabinets since 2002. They are providers of purpose-built, application-specific, server-appliance, and embedded systems platforms.

IRON has been hosting part of its infrastructure in the Azure cloud. IRON’s workloads include ERP, CRM, EDI, and in-house developed supply chain and other workflow management applications. IRON IT team had some issues with the database servers hosted on virtual machines that lead to degraded performance. IRON also wanted to leverage the MQTT protocol for one of their real-time application without hosting additional dedicated infrastructure. They had done some research and liked the number of options available on the AWS platform to ingest real-time data streams. They lacked an in-house AWS skillset and contacted Avahi to help with building POC. Upon further discussion, the Avahi team showed the ROI to IRON if they migrate all their workloads to AWS while improving the security and business continuity.

Challenges

IRON’s Azure infrastructure was provisioned and managed by multiple teams within the company. These teams had created additional subscriptions to host their workloads, creating a mix of conventions and practices when creating resource groups or configuring network architecture. Avahi worked with all these teams to understand their business needs and their ideal application hosting environments. From there, Avahi’s solutions architects worked backward to design an architecture that meets all stakeholders’ needs.

One of the IRON’s immediate needs was to enable its engineering team to provision resources on their own while keeping the cost in control and have proper governance in place. Avahi designed a solution based on Control Tower and Service Catalog. We put together a small POC for the stakeholders, which addressed all their concerns around enabling engineering team members to provision and manage their resources.

Avahi’s team realized the challenges they will run into when consolidating the disassociated environments and resources. We also anticipated the challenges we were going to face from the operational and logistics perspective.

Approach

The project required extreme coordination with all the parties and both business and technical stakeholders from various teams. Avahi utilized its Agile approach driven by the Jira project management portal to provide management visibility and meet the deadlines.

Architecture

Below is the architecture diagram of landing zone and network connectivity between on-premises data center and AWS environment.

We met the aggressive timelines of project execution from building the base environment to Go-Live in three months using a multi-phase approach.

Phase 1

The first phase was building a landing zone in AWS and lay down a proper account structure following the best practices. We utilized AWS Control Tower to build the landing zone. The first phase included design and automation for high-level tasks, including:

• Account Structure
• Create VPC and security groups
• Public/Private networks
• Firewall considerations
• Site to Site VPN using Transit Gateway
• Client VPNs (OpenVPN and IPSec)
• Cloudformation templates for IaC
• AWS Config setup
• CloudTrail setup
• IAM roles, users, and groups
• Active Directory integration
• Baseline documentation for future reference

Phase 2

The second phase was to lift and shift the virtual machines from Azure to AWS. We leveraged AWS Server Migration Service to migrate these resources. Avahi team addressed the licensing challenges for the Windows virtual machines and made the migration process seamless.

Phase 3

In the third phase, the Avahi team optimized the resources and migrated the databases from EC2 instances to the RDS service database. Auto-scaling rules were created for the web application to ensure high availability and optimize performance.