Episode Overview

Podcast: Marketecture: Get Smart. Fast.
Episode: Inside AWS’s RTB Fabric: Sergio Serra on Reinventing Ad Tech Infrastructure
Host: Ari Paparo
Guest: Sergio Serra, Product Manager Lead for RTB Fabric at AWS
Date: November 17, 2025

This episode dives into AWS’s new RTB Fabric—a purpose-built infrastructure solution designed for real-time bidding (RTB) in ad tech. Host Ari Paparo interviews product lead Sergio Serra to unpack how RTB Fabric addresses long-standing issues around latency, cost, and network complexity for SSPs, DSPs, and other industry players.

Key Discussion Points & Insights

1. What Is RTB Fabric?

• Purpose: RTB Fabric is AWS’s first purpose-built networking solution for the adtech industry, specifically tailored to real-time bidding workloads.
• Core Idea: Instead of handling infrastructure needs as an afterthought, RTB Fabric enables ad tech customers to communicate with each other in a low-latency, cost-efficient way.
• Abstraction: The system introduces “requester” (e.g., SSP, publisher) and “responder” (e.g., DSP, or even another SSP) roles, abstracting away the traditional infrastructure challenges.

  "RTB fabric is really a clear signal of AWS full commitment...Not just general proposed compute. We propose built infrastructure for the industry." (Sergio Serra, 00:45)

2. Latency, Cost, and Availability Zone Awareness

• Legacy Pain Points: Traditionally, even AWS customers have to use generic HTTP calls across regions/zones, incurring unpredictable egress costs and load balancing headaches.

• Improvements:

  • Cost: Eliminates standard egress and load balancing charges by bundling networking into a per-transaction fee model.
  • Availability Zone Awareness: Automatically ensures requests are routed within the same AZ for optimal latency.

  "RTB fabric allows you to really waive these two taxes...We bill for per billion requests or billion transactions." (Sergio Serra, 02:26)

• Real-World Example: Multi-AZ redundancy in Virginia is now routed deterministically for lower, predictable latency.

  “With RTB Fabric, [routing] becomes deterministic.” (Sergio Serra, 04:36)

3. Smart Load Balancing

• Traditional Complications: Customers usually need managed or self-hosted load balancers for TLS termination and fleet resolution—costly and operationally heavy.
• RTB Fabric Approach: DNS-based routing leverages customers’ autoscale groups or clusters, with smart, built-in load balancing, all bundled into the per-transaction charge.

  "We do that included in the service. There is no extra cost. It just came... in the cost per billion model." (Sergio Serra, 05:02)

4. Business Relationships & Onboarding

• Not a Market Facilitator: RTB Fabric does not bypass the need for legal/business/API agreements between parties.
• Network Setup Made Simple: Once the business agreements are in place, peering and actual network connection can be set up in as little as 10 minutes, vs. lengthy manual processes historically required.
• Open Internet Reach: Not limited to AWS-to-AWS traffic; can also facilitate external communications.

  "You just need 10 minutes in this case." (Sergio Serra, 06:55)
  “RTB Fabric...can actually go outside [AWS] very easily.” (Sergio Serra, 16:02)

5. Modular, Containerized Services

• Moving Beyond Monoliths: Encourages modular, containerized workloads instead of monolithic stacks.

• Network Modules:

  • Rate Limiter: Prevents QPS (queries per second) spikes from overwhelming DSP partners.
  • OpenRTB Filter: Rule-based filtering on OpenRTB attributes (e.g., by country, publisher format), with easy, granular configuration.
  • Error Masking: Masks backend errors (like 500s) as no-bids to prevent SSP circuit breakers from triggering unnecessarily.

  "The thing that is novel to me is...now you can run these modules tweaking the link." (Sergio Serra, 07:08)
  "We have some built in modules...Rate Limiter...OpenRTB filter...error masking." (Sergio Serra, 07:41)

6. Economics and Pricing Model

• Per-Transaction Charges: Charged per billion transactions, where both bid requests (requester) and responses (responder) count.
• No Hidden Fees: No standard AWS networking costs for covered payload size (P95 at 2KB); larger payloads incur a discounted per-GB charge.
• Tiered Pricing: Lower incremental costs for high-volume (over 2 trillion transactions).
• Support for External Transactions: Slightly higher pricing for transactions crossing into/outside AWS.

  "RTB fabric really allows you to reduce cost...by an order of magnitude." (Sergio Serra, 14:15)

7. Launch Partners and Ecosystem

• Early Partners: TripleLift, Viant, Ilmo, Mobile Fuse, Amazon Ads (APS & Amazon DSPs)
• Ongoing Growth: New partners joining regularly.

  "Some of them are Triple Lift, Viant, Ilmo and many more like Mobile Fuse. Of course we also work with Amazon ads..." (Sergio Serra, 15:26)

Memorable Quotes & Moments

• On Cost Alignment:

  "We were able to align cost on the unit economics of adtech. So we bill for per billion requests or billion transactions....Super cool in my mind." (Sergio Serra, 02:26)

• On built-in network modules:

  "Now the DSP can just say, hey, I don't want to have more than 1 million QPS from this specific link...we will throttle it for you." (Sergio Serra, 08:22)
  "You still have to deserialize the request, usually, right? Even if you don't care about that, you pay a cost for deserializing that. In our case, we do it for you..." (Sergio Serra, 08:54)

• On error masking:

  "We temporarily can mask that 500 to a 204, which is a no bid...super fast because they execute in the network." (Sergio Serra, 09:24)

• On competitive advantage:

  "The size of AWS allows us to, of course, be very cost efficient." (Sergio Serra, 16:15)

• On RTB Fabric’s animal spirit:

  "I would say it's a lion because it's gonna really break the industry on that extent." (Sergio Serra, 16:28)

Notable Timestamps

• 00:45 – Sergio Serra explains what RTB Fabric is.
• 02:26 – How RTB Fabric lowers costs and provides per-transaction billing.
• 04:36 – Availability Zone awareness and deterministic routing.
• 05:02 – How Fabric handles load balancing with DNS-based routing.
• 06:55 – Simplicity of onboarding after business relationships established.
• 07:41 – Introduction to built-in network modules (Rate limiter, OpenRTB filter, Error masking).
• 10:40 – Deep dive on modular filtering capabilities.
• 11:34 – Detailed explanation of the economic model and fee structure.
• 14:15 – Concrete cost reduction examples and networking fee elimination.
• 15:26 – Launch partners announced.
• 16:02–16:37 – Lightning round: misconceptions, competitive advantages, and "if Fabric were an animal..."

Conclusion

This episode provides a comprehensive overview of AWS’s RTB Fabric—its technical and commercial innovations, the industry pain points it addresses, and new operational paradigms for both SSPs and DSPs. Sergio Serra offers transparent insight into why AWS built this product, who it serves, how it works, and the practical financial advantages for early adopters.

Listeners walk away with a strong understanding of both the technology and the rationale behind AWS RTB Fabric, directly from the product’s guiding hand.