Datashapes DevOps
Highly available multi-tenant production grade
SaaS applications on Azure platform
- Industry : Internet of Things
Service Offered
Cloud Ops
Overview
Celestial deployed a multi-tenant SaaS application to the production-grade Azure cloud Infrastructure for our client, Datashapes.
This same concept can be extended to other SaaS applications that may have similar use-case/requirements. The idea was to plan, design and implement an architecture that is highly available, scalable and secure.
Dissecting Problem
Know the Client
A San Francisco Bay Area company, that is commercializing the patented, full-stack, market-ready DataShapes AI platform, in an enterprise-grade SaaS platform
Business Requirement
- Architect a cloud infrastructure design that can host and support SaaS applications and their tenants.
- To build a scalable, highly available, and yet secure platform.
- Support multi-tenant architecture.
- The system should support auto-scaling.
Challenges
The main challenge was infrastructure resource provisioning. To avoid the wastage of resources, the resource allocation should be in such a way that, it should support multi-tenancy without having too many overheads in terms of resources
Celestial’s Solution
Celestial DevOps team assisted in planning, designing, and implementing the architecture that could support multi-tenant functionality of the SaaS application. We designed a solution based on the Azure ecosystem in allocating dedicated resources for each new tenant onboarded.
MULTI-TENANCY
Dedicated and isolated set of resources for each tenant onboarded
SCALABILITY
The application resource sets should be able to auto-scale themselves to improve performance
HIGH-AVAILABILITY
Offers a high level of operational performance and quality over a relevant time period
SECURITY
Access of the DevOps/Admin team to the infrastructure should be authenticated via Azure AD
Technology Used :
Outcomes
- Implemented a multi-tenant architecture
- Automated provisioning of Kubernetes objects
- Azure AD integration