RTB Latency Budget: QPS, Sizing and Measurement

RTB latency is an end-to-end budget. Network travel, request handling, data lookups, bidding and response delivery must fit inside the exchange deadline. QPS describes load, while tail latency shows how that load affects slower requests. Size the system from measured behavior and test failure conditions before committing capacity.

What QPS Really Means in AdTech

QPS vs Fill Rate: The Critical Relationship

Example calculation:

How Latency Impacts Revenue

The Revenue Cost of Latency

* Based on industry benchmarks and Google research on latency impact

Network, Storage, and Compute Bottlenecks

Performance bottlenecks hide in three layers: network (connectivity), storage (database/disk), and compute (CPU/memory). Identifying the constraint is the first step to optimization.

┌─────────────────────────────────────────────────────────┐
│  PERFORMANCE BOTTLENECK IDENTIFICATION                  │
├─────────────────────────────────────────────────────────┤
│                                                         │
│  START: High latency or low QPS detected               │
│                                                         │
│  1. Check Network                                       │
│     ├─ Packet loss? → CDN or direct connection         │
│     ├─ High ping? → Geographic routing issue           │
│     └─ OK? → Continue to storage                       │
│                                                         │
│  2. Check Storage                                       │
│     ├─ Slow queries? → Add indexes or optimize SQL     │
│     ├─ High IOPS? → Upgrade to SSD/NVMe                │
│     ├─ Replication lag? → Scale read replicas          │
│     └─ OK? → Continue to compute                       │
│                                                         │
│  3. Check Compute                                       │
│     ├─ High CPU? → Optimize algorithms or scale out    │
│     ├─ Memory swapping? → Increase RAM or reduce load  │
│     ├─ Cache misses? → Tune cache size or TTL          │
│     └─ OK? → Check application logic                   │
│                                                         │
│  4. Application Logic                                   │
│     ├─ Complex loops? → Refactor to O(1) lookups       │
│     ├─ Blocking I/O? → Switch to async/non-blocking    │
│     └─ External API calls? → Cache or batch            │
│                                                         │
└─────────────────────────────────────────────────────────┘

Network Bottlenecks

• Fix: Edge nodes in major markets, optimized routing, CDN for static assets

Storage Bottlenecks

• Fix: Upgrade to SSD/NVMe, optimize queries, add composite indexes, use read replicas

Compute Bottlenecks

• CPU saturation: Reduces QPS capacity, increases latency variability
• Memory bandwidth: Critical for in-memory caching and large datasets
• Fix: Horizontal scaling, algorithm optimization, GPU for ML-based bidding

15-Point Performance Audit Checklist

Use this checklist to audit your RTB infrastructure performance. Each item includes target benchmarks and common fixes.

Identify peak traffic patterns (hourly, daily, seasonal)

Audit database query performance (slow query log)

Low hit rate = wasted cache capacity or wrong TTL

Test network latency to major exchanges

Use traceroute and ping tests from production servers

Monitor disk I/O wait times

High iowait = storage bottleneck, consider SSD upgrade

Check load balancer distribution

Uneven load = routing issue or unhealthy node

Review bidding algorithm complexity (O(n) vs O(n²))

Test disaster recovery failover time

Measure data replication lag

Audit logging overhead

Test geographic redundancy

Simulate datacenter failure, does traffic failover correctly?

Review rate limiting policies

Protect against DDoS without blocking legitimate traffic

Get Performance Audit Template

Download our comprehensive Excel template with pre-configured formulas, benchmark ranges, and automated health scoring for your RTB infrastructure.

Key Takeaways

• ✓ Bottlenecks hide in network, storage, or compute, systematic diagnosis required
• ✓ Multi-tier caching (in-memory, Redis, database) is critical for scale
• ✓ Predictive monitoring and autoscaling prevent performance degradation before it happens

• bidder capacity planner
• prebid server self hosting guide
• rtb architecture

Read the rtb architecture guide for the broader decision framework.

Is Prebid Server free? The software is free under the Apache-2.0 license for both implementations (prebid-server README; prebid-server-java README). Running it is not: Prebid.org says it "is not a turn-key solution" and that the host must build its own data pipeline for accounts and stored requests (Prebid Server hosting). The worksheet above prices the infrastructure.

Should I choose Prebid Server Go or Java? Prebid.org says to pick the language your team prefers if the features and adapters you need exist in both (PBS-Go versions page). The feature index lists 16 Java-only features and 4 Go-only ones; US state privacy (USGen module), the circuit breaker and per-account stored requests are Java-only, while geo-fencing and legacy schain locations are Go-only (feature index). Microsoft runs its commercial service on Go (Microsoft Learn, 21 October 2025).

How much does it cost to run Prebid Server on AWS? It depends on request rate, regions and egress, none of which Prebid.org benchmarks. The sourced unit prices are $0.145 per hour for a c7g.xlarge and $0.082 per hour for an m7g.large on demand, $0.0225 per hour plus $0.008 per LCU-hour for an Application Load Balancer, and $0.1752 per node-hour for a cache.r7g.large in AWS's own example (Vantage; AWS ELB pricing; AWS ElastiCache pricing).

How many bidders does Prebid Server support? Prebid.org's overview says "230+" server-side adapters; the bidder list page carries 394 entries, some of which are aliases or duplicates (Prebid Server overview; bidder list). Each adapter declares its endpoint, GVL ID, media types and user-sync method in a bidder-info file (adding a new bidder, Go).

Do I need Prebid Server to get AdX bids after the Google ruling? Not necessarily, but it is one of the two court-ordered paths. Google must build an API so Prebid can solicit real-time AdX bids and a DFP-to-Prebid-Server link (Google's estimate 12 to 15 months from the Effective Date), and separately must send AdX bids to rival ad servers (6 or 12 months, disputed) (Memorandum Opinion, Dkt. 1858, p. 58; PPC Land, 6 October 2026). A Prebid Server you host, or one a managed host runs for you, is where the first path lands; open-web display only.

1. GitHub, prebid/prebid-server, README and Releases (accessed 17 October 2026). https://github.com/prebid/prebid-server and https://github.com/prebid/prebid-server/releases
2. GitHub, prebid/prebid-server-java, README and Releases (accessed 17 October 2026). https://github.com/prebid/prebid-server-java and https://github.com/prebid/prebid-server-java/releases
3. GitHub, prebid/prebid-server, docs/developers/configuration.md (accessed 17 October 2026). https://github.com/prebid/prebid-server/blob/master/docs/developers/configuration.md
4. GitHub, prebid/prebid-server, config/config.go defaults (accessed 17 October 2026). https://raw.githubusercontent.com/prebid/prebid-server/master/config/config.go
5. GitHub, prebid/prebid-server-java, docs/config-app.md (accessed 17 October 2026). https://github.com/prebid/prebid-server-java/blob/master/docs/config-app.md
6. pkg.go.dev, github.com/prebid/prebid-server/v4 versions, with GO-2026-6139 advisory (accessed 17 October 2026). https://pkg.go.dev/github.com/prebid/prebid-server/v4?tab=versions
7. Prebid.org, "Prebid Server overview" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/overview/prebid-server-overview.html
8. Prebid.org, "Prebid Server hosting" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/hosting/pbs-hosting.html
9. Prebid.org, "Prebid Server feature index" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/features/pbs-feature-idx.html
10. Prebid.org, "Prebid Server bidder list" (accessed 17 October 2026). https://docs.prebid.org/dev-docs/pbs-bidders.html
11. Prebid.org, "Adding a new bidder (PBS-Go)" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/developers/add-new-bidder-go.html
12. Prebid.org, "Stored requests (PBS-Go)" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/features/pbs-storedreqs-go.html
13. Prebid.org, "Prebid Server cookie sync" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/developers/pbs-cookie-sync.html
14. Prebid.org, "Prebid Server privacy features" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/features/pbs-privacy.html
15. Prebid.org, "Prebid Server price floors" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/features/pbs-floors.html
16. Prebid.org, "Prebid Server adapter (s2sConfig)" (accessed 17 October 2026). https://docs.prebid.org/dev-docs/modules/prebidServer.html
17. Prebid.org, "PBS-Go versions" (accessed 17 October 2026). https://docs.prebid.org/prebid-server/versions/pbs-versions-go.html
18. Prebid.org, "Managed services" (accessed 17 October 2026). https://prebid.org/managed-services/
19. Microsoft Learn, "Prebid Server Premium" (21 October 2025). https://learn.microsoft.com/en-us/xandr/monetize/prebid-server-premium
20. Vantage (EC2Instances.info), c7g.xlarge and m7g.large (updated 17 October 2026). https://instances.vantage.sh/aws/ec2/c7g.xlarge and https://instances.vantage.sh/aws/ec2/m7g.large
21. AWS, Elastic Load Balancing pricing (accessed 17 October 2026). https://aws.amazon.com/elasticloadbalancing/pricing/
22. AWS, ElastiCache pricing (accessed 17 October 2026). https://aws.amazon.com/elasticache/pricing/
23. AWS, EC2 on-demand pricing, data transfer section (accessed 17 October 2026). https://aws.amazon.com/ec2/pricing/on-demand/
24. United States v. Google LLC, No. 1:23-cv-00108 (E.D. Va.), Memorandum Opinion, Dkt. 1858 (2 September 2026). https://www.courthousenews.com/wp-content/uploads/2026/09/google-judge-memo.pdf
25. PPC Land, "DOJ seeks 6 months, Google 12, to open its ad exchange to rivals" (6 October 2026). https://ppc.land/doj-seeks-6-months-google-12-to-open-its-ad-exchange-to-rivals/

• October 18, 2026: aligned the approved search title and heading; retained the original publication date.

• October 17, 2026: updated search phrasing, source-linked reference context, author and dates. Withheld numerical claims without inline sources.