Forem: aaronkao

The Promise of the Cloud, Unfulfilled

aaronkao — Fri, 02 Aug 2024 15:57:21 +0000

When your company moved to the cloud, it was supposed to supercharge innovation while slashing costs. Unlimited scalability and agility were within reach. But for many, that potential remains frustratingly untapped.

Instead of accelerating development, managing infrastructure has become a bottleneck. Your DevOps and platform engineering teams are mired in:

Clunky, manual processes - spinning up a simple test environment takes days
Painfully slow rollouts - deploying updates across regions drags on for weeks
Overwhelming complexity - the combinations of cloud services seem infinite

There has to be a better way to truly harness the cloud and turn it into your secret weapon. That's where Pulumi comes in.

Pulumi: The 10x Better Approach to Infrastructure as Code

Pulumi is an open source infrastructure as code (IaC) platform that enables you to provision and manage your cloud with the same languages and tools you use to build applications. This unlocks powerful benefits:

Familiar languages and tools: Use Python, TypeScript, Go, or any language you love. Tap into AI copilots, IDEs, testing frameworks, and CI/CD.

Built for modern architectures: Easily automate and scale microservices, containers, serverless, and more. Apply software engineering to tame the complexity.

AI-powered user experience: Semantic understanding of your cloud combined with powerful language models provides unique insights and automations.

Pulumi takes a fundamentally different approach that delivers 10x better productivity and time-to-value compared to other IaC tools. But what does that mean for you?

Compelling Reasons to Make the Switch

Adopting Pulumi, trusted by hundreds of thousands of developers, can be game-changing:

Accelerate deployments. Provision environments in minutes instead of days. Roll out updates in hours instead of weeks. Deliver value faster.
Increase productivity. Use best-in-class languages and tooling to manage infrastructure. No more fighting clunky domain-specific languages.
Automate at infinite scale. Use software engineering to tame the mind-boggling combinations of cloud services. 10x your infrastructure-to-engineer ratio.
Open source stability. Avoid the uncertainty of tools altering licenses and tying open source to commercial services. Get dependable open source.

Pulumi customers routinely see their productivity double, provisioning times compress 99%, and infrastructure management accelerate 10x. But Pulumi shines for specific use cases too.

Why Pulumi for Internal Developer Platforms?

Building an Internal Developer Platform (IDP) to empower your developers requires:

Developer Control Plane - Curated experiences and self-service interfaces
Integration & Delivery - Automation to provision, test, and deploy all aspects of infrastructure
Monitoring & Logging - Observability to maintain operational control and optimize costs
Security & Identity - Guardrails to enforce policies and control access
Diverse Resources - Support for Kubernetes, serverless, AI/ML, hybrid cloud, and more

Pulumi provides the building blocks for each layer, enabling you to craft an IDP that bridges the gap between your developers and the cloud. Abstract away complexity while maintaining security and compliance.

Why Pulumi for AI Workloads?

Deploying AI applications requires provisioning and gluing together many layers:

Model training infrastructure
Data transformation pipelines
Backend services
Web frontends

Pulumi allows you to define your entire AI stack, from machine learning frameworks down to cloud services, as a simple Python library. Go from local development to cloud production environments with ease.

Why Pulumi Over Alternatives?

There are many infrastructure tools out there, but Pulumi delivers an unmatched combination of productivity and power:

Pulumi vs Clickops

Manually provisioning prod infra = bad idea. No repeatability or consistency. Errors lead to downtime or worse. Automate with IaC.

Pulumi vs Terraform

Both are IaC, but Pulumi enables far greater productivity, scalability, and collaboration:

Use real languages + tools, not proprietary DSLs
Apply software engineering to tame complexity
Empower self-service with IaC building blocks

Pulumi vs AWS CDK

Both use code for IaC, but Pulumi provides:

Multi-cloud with no lock-in, not just AWS
Direct execution, not transpiling to CloudFormation
Versatile testing, including fast in-memory mocking

Pulumi vs IDP Services

3rd party IDPs provide simple portals but lack Pulumi's:

Infrastructure as code foundation
Powerful customization through software engineering
Enterprise-grade secrets management and policy enforcement

Making the Switch is Easy

Adopting Pulumi doesn't have to be daunting. We've helped thousands of customers do it:

Ask questions on our Community Slack
Use our self-serve tools to import existing infrastructure
Get expert help from our team to plan and execute your migration

The cloud's promise is within your grasp. You just need the right tools to seize it.

Pulumi can be the bridge between your teams and the cloud. The key to turning infrastructure from a bottleneck into a launchpad.
So why wait? Join the thousands of companies achieving 10x greater velocity in shipping products.

Header image credit: ESO/M. Kornmesser

Building High-scale AI Backends

aaronkao — Tue, 05 Dec 2023 22:53:44 +0000

What does it take to go from an idea in a notebook to an application handling real-world traffic?

The Pulumi and Pinecone teams worked together to build a reference architecture that teaches you how to scale AI apps in production. We tested it with batches of 10k to 1M records. The reference architecture demonstrates microservices scaling, data processing pipelines, infrastructure segmentation through networking and security groups, and UI and database synchronizations.

Architecture Details

On the infrastructure side, the reference architecture uses a Pinecone index as a vector store, a queue to fan out work, networking and security groups to separate the infrastructure, ECS services for the frontend and backend microservices, and autoscaling configured to expand the worker pool up and down elastically in response to system load.

On the application side, the reference architecture launches a semantic search interface over a Postgres database of product records, leveraging Pinecone’s vector database for queries and instant index updates.

The reference architecture is built with the best practices for how to use AWS and Pinecone. It is designed for production, and it is easily modifiable to fit most use cases.

Try it out: https://github.com/pinecone-io/aws-reference-architecture-pulumi

Additional Resources

Secrets and configuration management made easy

aaronkao — Fri, 20 Oct 2023 19:06:28 +0000

Pulumi ESC is a new product from Pulumi that manages and tames secrets and configuration complexity across all of your cloud infrastructure and application environments. Pulumi ESC introduces a new category of configuration-as-code product, motivated by our experience working with hundreds of Pulumi IaC customers to address their needs in managing secrets and configuration at scale within their Pulumi infrastructure and across other cloud applications and infrastructure projects.

Infrastructure as Code best practices

aaronkao — Wed, 20 Sep 2023 22:15:54 +0000

We frequently get questions from our users on what are some of the best practices for infrastructure as code: how to organize the code, how to structure projects, how to manage access controls, how to do cross stack references, how to build control planes, etc. Most of this information was dispersed across documentation, Slack conversations, and blog posts or just not written down. So we created a reference architecture that has all these best practices built in and then walked through it all in a series of blog posts.

Follow along here:

Deploying Hugging Face models on Amazon SageMaker using infrastructure as code

aaronkao — Tue, 19 Sep 2023 22:00:32 +0000

Here is a quick template and tutorial for bootstrapping Amazon SageMaker to run a LlaMa 2 model from Hugging Face. Everything is deployed as code (Python); no manual tweaking in the SageMaker console.

www.pulumi.com/blog/mlops-huggingface-llm-aws-sagemaker-python/