Dapr's Road Ahead: GenAI APIs, Distributed Scheduling at Scale and What It Means - Kubecon Session

In this maintainer track presentation, Yaron Schneider covers the latest developments and updates of the Dapr project looking into 2025, focusing on how Dapr is adding APIs to abstract the complexities of interacting with LLM models at scale, a new distributed scheduling API and workflow engine that can serve millions of activities per second and how Dapr can be used by platform teams to provide golden paths for interacting with the underlying infrastructure.

Beyond DevOps: The Rise Dapr in Platform Engineering

Unlocking Dapr's Potential in Advanced Application Engineering

The world of cloud-native development continues to evolve, and the Dapr project has become a cornerstone for many organizations striving to simplify complex infrastructure interactions. As one of the Dapr core maintainers, Yaron Schneider recently shared the project's journey, growth, and its alignment with platform engineering—a strategic discipline reshaping how enterprises manage and scale their applications.

Dapr: A Snapshot of Adoption

Since being donated to the Cloud Native Computing Foundation (CNCF), Dapr has seen remarkable growth. Over the past year, its SDK downloads surged: Python grew by almost 200%, JavaScript by 109%, and .NET has exceeded 1.5 million downloads per major version. These statistics reflect only part of the picture, as they don’t account for Dapr's adoption via HTTP or gRPC. This adoption underscores its appeal across languages and platforms.

Some impressive use cases include:

Grafana Labs: Leveraging Dapr for their software supply chain.
HDFC Bank: Processing 750 million transactions monthly.
Vonage: Scaling Dapr across 100 Kubernetes clusters on three clouds and on-premises.
Wattswater: Utilizing Dapr workflows and actors at scale.

Such use cases highlight Dapr’s flexibility and scalability across industries and geographies.

The Shift to Platform Engineering

Initially championed by individual developer teams, Dapr is now primarily adopted by platform engineering teams. This shift emphasizes centralized platform strategies that provide reusable building blocks for development teams, enabling them to focus on business logic rather than infrastructure concerns.

Platform teams encapsulate infrastructure complexities, offering APIs that integrate monitoring, provisioning, and container orchestration seamlessly with application code. Dapr’s APIs extend this abstraction layer, ensuring developers can access:

Reliable messaging: PubSub support with pluggable backends.
State management: Simplifying database and cache interactions.
Workflows and actors: Supporting sequential and parallel task execution.
AI integrations: Providing APIs for seamless interaction with large language models (LLMs).

Key Innovations in Dapr 1.15

1. Streaming PubSubDapr 1.15 introduces a game-changing feature: streaming PubSub. Unlike traditional push-based PubSub models, this allows developers to pull messages, providing better control and backpressure management. This is especially valuable for high-performance use cases where latency and reliability are critical.

2. Workflows at ScaleWith the latest updates, Dapr workflows can scale to millions of concurrent executions, making them ideal for data processing and AI workloads. Whether performing sequential tasks or managing reconciliation loops akin to Kubernetes controllers, workflows empower developers to design complex systems with ease.

3. Conversational API for AIDapr now supports interactions with major AI platforms, including AWS Bedrock, Hugging Face, OpenAI, and Anthropic. The API simplifies secure, efficient communication with LLMs and incorporates features like sensitive data obfuscation and prompt caching—ensuring both cost-effectiveness and compliance.

Enterprise-Grade Features and Roadmap

The Dapr team is committed to enhancing enterprise capabilities, as evidenced by their long-term roadmap:

SQL and Blob Storage APIs: Expanding beyond key-value stores to support advanced queries and semi-structured data.
Delayed PubSub: Scheduling messages for future delivery.
Authorization API: Introducing a robust RBAC system within the Dapr sidecar.
Improved Job Scheduling: Enabling distributed cron-like functionality.
Bulk and Batch Processing: Enhancing high-throughput scenarios with documentation and SDK support.

These initiatives aim to make Dapr a one-stop platform for cloud-native development, fostering collaboration across diverse application teams.

Dapr and the Future of Platform Engineering

Dapr’s evolution reflects the growing importance of platform engineering. By abstracting infrastructure complexities and offering best practices in a box, it equips organizations to focus on innovation and scalability. As enterprises adopt Dapr as a core component of their platforms, the lines between developer enablement and infrastructure management blur, paving the way for a new era in software development.

Dapr’s journey is a testament to community-driven innovation. As Yaron Schneider emphasized, the contributors—developers, maintainers, and users—are the lifeblood of this project. With its recent CNCF graduation, Dapr is poised to continue transforming cloud-native ecosystems for years to come.

‍