An Azure Hybrid Infrastructure

Client

PSI CRO is a full-service global Contract Research Organization specializing in delivering clinical projects across multiple therapeutic areas.

Challenge

To enhance the technology infrastructure and streamline PSI operations, the company required an internal network in the cloud with the ability to access on-premises services. The main challenges leading to this request included:

• The geographical distribution of project teams requiring a unified solution to foster efficient collaboration;
• The necessity to enhance the responsiveness of business applications;
• Cost optimization for the company’s Data/AI/ML infrastructure, as traditional on-premises data centers often incur high costs;
• The need for integration between existing applications to ensure a smooth and uninterrupted data flow across the entire organization.

Thus, these factors led the DataArt experts to integrate Azure cloud-based resources with the existing on-premises infrastructure, creating a hybrid infrastructure model for the client. This solution allowed the client to harness the advantages of Azure Cloud while leveraging their on-premises resources.

DataArt was a fantastic partner in our project. Together, we were able to take our inexperience with Azure Cloud and turn it into an opportunity to develop and implement an entire DevOps methodology from scratch. In large part, that was possible because of DataArt’s ability to listen and hear the client’s needs as well as the cohesiveness of the joint team.

Kirill Soldatov

Head Process Improvement, PSI

Solution

Several services were provided by DataArt in this project’s scope, including:

• Developing DevOps practice inside PSI from scratch: The client had no prior experience in Azure cloud, so DataArt’s team developed and implemented the entire DevOps methodology from scratch. The existing on-premises infrastructure was extended to the cloud, enabling the client to leverage the benefits of Azure cloud computing while still utilizing their on-premises resources.
• Establishing Azure cloud governance: As part of Azure cloud governance, the team defined guidelines and standards for accessing the cloud, established naming conventions, determined artifact storage practices, implemented cloud governance policies, and more.
• Fixing PSI hybrid network implementation (on-prem + Azure cloud) and improving security: The issues or inefficiencies in the existing network setup that combines on-premises infrastructure with Azure have been resolved. The focus was on enhancing network performance, connectivity, and overall security.
• Usage of DNS services in the network setup: Since all resources were deployed without external access, Azure DNS Forwarder has been set up for seamless integration between the local customer network and Azure cloud network.
• Deployment with Azure Kubernetes Service (AKS): DataArt's solution utilized AKS and intelligent scaling based on a queue system. This approach accelerated computations while optimizing the total cost of infrastructure ownership.

Highlights

• Stable and secure hybrid network with Azure
• Faster and more precise calculations
• Cost efficiency for the scalable Data/AI/ML infrastructure
• Unified network for geographically distributed users
• Seamless data transfer and workload migration between environments

Technologies and Tools