What was proposed:

With the advent of the COVID-19 pandemic, the company decided to start a path of complete migration of its technology operations to the public cloud. Previously, the entire IT infrastructure was dispersed in multiple locations, with decentralized management, making the challenge of growth even greater. With the choice of AWS and Valcann, the entire project was conducted with the scope of:

• Map and model the entire technology architecture;

• Propose technical migration scenarios;

• Conduct a capacity planning and infrastructure dimensioning study;

• Optimize systems architectures and business applications;

• Migrate Windows Servers and databases (SQL Server, Oracle, MySQL) and other business applications;

• Predict backup and disaster recovery strategy;

• Migrate TOTVS Withor business application;

• Provide specialized support in cloud computing architecture and technology operation.

Technical enablement

The main challenge of the ABC Distribuidora project was the validation process of all business units. It is a distributor of graphic components, spread throughout Brazil. There was a lot of customer concern to ensure the least impact cut from the previous datacenter environment to AWS. For this, the validation and approval process went through a wide cycle of functional and non-functional tests, validating operations both with key users and performance aspects (through load tests). Another significant component in this project was cost. There was a maximum cost markup which, if exceeded, would make the project - and, consequently, the entire migration - unfeasible. In this sense, the load tests were essential to define whether the sizing performed previously would be sufficient to support the real demand of the customer.

AWS Services deployed: 

• Amazon EC2, Amazon RDS, Amazon S3, Amazon CloudWatch, Amazon EFS, Amazon GuarDuty, AWS CloudTrail, CloudEndure.

Customer issues and concerns that have been overcome:

1. Operational excellence

1.1. Implementation of continuous monitoring with Amazon CloudWatch;

1.2. Integrated connectivity with high available multi zone (Multi-AZ);

2. Security

2.1. Implementation of Amazon GuardDuty for intelligent monitoring of infrastructure perimeter;

2.2. Restriction of external access to the environment, with segregation between private and public subnets for back and front end, respectively;

2.3. Access to the AWS environment for VPN-only administration;

2.4. Encryption of data at rest and in transit.

3. Reliability

3.1. Multiple instance nodes, making it possible to balance reading and writing operations;

3.2. Minimal failover between nodes;

3.3. Using Amazon Data Lifecycle Manager to implement snapshot-level backups.

3.4. Automatic backups of the database layer;

4. Excellence in performance

4.1. Calculation of expected environment demand for rightsizing.

5. Cost optimization

5.1. The monitored environment, with a continuous flow of monitoring for rightsizing.

Third-party solutions applied:

• • TSPlus: Remote Terminal Service;

  • FreeBSD pfSense: VPN connectivity;

  • Hashicorp Terraform: Infrasctructure as Code;

  • Hashicorp Vault: Secrets and key management service;

  • Oracle Database Standard Edition Two: Oracle Database on Amazon EC2 for business verticals applications.

Project schedule:

Project Start: January, 2021

Projet End: March, 2021

Key Results:

• • AWS managed infrastructure service;

  • Professional managed services provided;

  • Integration of 11 different business units in a centralized infrastructure in AWS Cloud;

  • Performance improvement of TOTVS application;

  • Application virtualization, improving application update time from 3 hours to update all clients to 20 minutes.

Lessons learned:

• • Carrying out load tests before going into production, makes it possible to validate if the capacity planning was adequate;

  • It is important to map the on premises structures of clients that have multiple business units, in order to minimize errors due to non-standardization of architecture.