Hybrid Cloud-Native Networking in Enterprise - Some Assembly Required

Hybrid Cloud-Native Networking in Enterprise - Some Assembly Required

The presentation, delivered by Louis Ryan, explores the evolving challenges of networking within enterprises, particularly in the context of hybrid cloud-native environments. It highlights the complexities and adjustments required as businesses strive to integrate multi-cloud strategies, containerization, and modern networking infrastructure.

Key Points Discussed:

• Network Functionality and Evolution: Networks are often too primitive, focusing more on moving data at high speeds rather than offering useful features for applications.
• Application vs. Network Prioritization: There's a call to make networks more beneficial to applications rather than merely serving the needs of network administrators.
• Networking Identity: Emphasizes the significance of introducing verifiable identities in networking to enhance security and simplify policy management.
• Policy Composability: Highlights the need for composable policies to ensure networks and their components can integrate seamlessly and remain flexible.
• Proxy and Policy Patterns: Discusses common network management patterns such as using large proxies and centralized policy stores, their benefits, and potential drawbacks.
• Infrastructure Repeatability: Stresses the importance of developing networks that provide consistent controls across diverse infrastructure setups, aided by commodity infrastructure like open-source tools.

This presentation combines technical insights with philosophical perspectives, urging a shift in how networking is approached in enterprise environments to better support application development and business objectives.

This is the end of the AI-generated content.

Delivering a 'useful' network in the typical enterprise has always been challenging. Whether it's the rush to containerization, business goals around accelerated value delivery or regulatory pressure to be multi-cloud or even resurrect your own data centers, that job keeps getting harder. How should we evaluate the tools we use and how do we want them to compose to be useful?