Summary
This document is a technical report that outlines the architecture and deployment models for AI and machine learning applications. It discusses the complexities introduced by generative AI, emphasizing the need for multi-modal app experiences and the influence of 'data gravity' on application and model placement. The report details four deployment models: SaaS AI, Cloud-Hosted AI, Self-Hosted AI, and Edge-Hosted AI, each with distinct characteristics and benefits. Additionally, it presents considerations for the AI ecosystem, including security risks associated with large language models (LLMs) and application delivery challenges in hybrid multicloud environments. The report also introduces seven AI building blocks, which include components such as fine-tuning services, training services, and inference services. Each building block is defined, explaining its role in the overall architecture and how it contributes to the deployment and functionality of AI solutions.
