Optimizing network efficiency & strengthening multicloud connectivity (MAE308)

Optimizing Network Efficiency and Strengthening Multicloud Connectivity

New York Times' Cloud Journey

  • New York Times started their cloud journey by dividing their workloads between on-premises and two public cloud providers.
  • A few years later, they decided to migrate 80-90% of their workloads from the other public cloud provider to AWS.
  • This migration process is still ongoing, and the presenters will share the key stories and use cases from this journey.

Developing an Internal Developer Platform

  • The goal was to provide a consistent experience and accelerate the development lifecycle for engineers across the organization.
  • Key elements:
    • Standardization: Ensure a cohesive and consistent experience for engineers.
    • Efficiency: Eliminate repetitive tasks and abstract infrastructure complexity.
    • Integration: Integrate multiple systems and cloud providers into a cohesive platform.
    • Scaling: Enable scaling from 0 to 100 with reduced overhead.
    • Visibility: Provide performance and cost insights to enable continuous improvement.

Networking Challenges and Solutions

Networking Cost Optimization

  • Network Address Translation (NAT) can become expensive at scale, especially when moving data between cloud providers.
    • Solution: Implement custom NAT instances instead of relying solely on AWS NAT Gateways.
  • Fixed networking costs (e.g., multiple NAT Gateways) can become unnecessary overhead as the platform scales.
    • Solution: Centralize routing to eliminate unnecessary fixed costs.

Multicloud Connectivity

  • Connecting workloads across cloud providers through the public internet can be costly and expose security risks.
  • Key requirements:
    • Predominantly US East to US East traffic
    • Require at least 10 Gbps of connectivity between cloud providers
  • Solutions:
    • Leverage existing on-premises Direct Connect connectivity to connect to cloud providers.
    • Explore Network-as-a-Service providers with pre-built infrastructure in major POPs.

Architectural Considerations

  • Adopted a hub-and-spoke model for each cloud provider, with the networking team hosting the hub account.
  • Ensured high availability and redundancy (minimum 4 NIC connections) in the design.
  • Implemented least-privileged access and self-service capabilities for application teams.
  • Utilized BGP techniques like prefix-length, AS-path prepending, and BFD for failover to optimize routing.

Key Outcomes and Takeaways

  • Achieved a 60% reduction in transfer costs by eliminating cross-cloud egress.
  • Improved application performance with 20-40% lower HTTP latency between cloud providers.
  • Reduced security exposure by 50% by minimizing external-facing attack surfaces.
  • Enabled seamless migration of services between cloud providers by providing a consistent networking foundation.
  • Delivered a self-service platform for application teams to innovate and scale more efficiently.

Your Digital Journey deserves a great story.

Build one with us.

Cookies Icon

These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.

If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.

Talk to us