1 min read
Next-Generation Phi Models: Advancing AI for All Devices
Microsoft has officially released the next generation of Phi models, a set of Small Language Models (SLMs) designed for efficiency, speed, and...
Okay, so we are ready to go the modernization route, but what’s the benefit in the end? We’ll break that down based on the landing spot, but we often see some hybrid models where applications span two or more service models. Your optimal path to the cloud will depend on your environment, assets, investments, and existing technical debt. Generally speaking, abstraction and consumption costs go down as we move from the left to the right. However, our degree of control of the environment also diminishes.
Cloud Service Models
Infrastructure as a Service (IaaS)
Benefits
-
Avoid capex and hardware support costs
-
Scalability, redundancy, HA configurable
Common Scenarios
-
3rd Party Apps not suitable for containers
-
Apps with limited lifespan, thus no ROI to alter
-
Temporary landing spot to address hardware end of life
-
Applications that require control or interaction with the OS
Containers as a Service (CaaS)
Benefits
-
More easily scale horizontally and vertically when architected properly
-
Dynamic scaling up/down can improve consumption costs and performance
Common Scenarios
-
3rd Party Apps or components with Kubernetes or Docker versions already available
-
Legacy code or components that can’t be deployed to PaaS
-
Move to PaaS or FaaS to involved/costly
Platform as a Service (PaaS)
Benefits
-
More focus on code and less on deployments
-
Very little concerns with scaling, cloud vendor handles it
-
Pay for consumption (cpu, memory, storage)
Common Scenarios
-
Apps with simple data access layer can generally be retrofitted to Cloud SQL (or similar PaaS DB)
-
Apps that need to manage session and/or state
Functions as a Service (FaaS)
Benefits
-
Only focus on code
-
Generally enforces a simple event & services based architecture, increasing supportability
-
Scalability is generally not a concern at all
Common Scenarios
-
Event based architectures are most common
-
Apps (or portions of apps) based on Microservices
-
Stateless services
-
Applications that require control or interaction with the OS
1 min read
DeepSeek R1 Now Available in Azure
Admin : Jan 7, 2026 12:14:27 PM
The latest frontier in AI reasoning models, DeepSeek R1, is now available in Azure AI Foundry and GitHub. Designed to provide developers with...
2 min read
Azure AI Studio is Expanding: New Models and Enhanced Tooling Support Now Available
Admin : Jan 7, 2026 12:13:57 PM
Azure AI Studio (NOT to be confused with Copilot Studio) is undergoing significant expansion, welcoming a host of new models while bolstering its...