AWS wellarchitected: Restructure ADVPERF01 with capability intent and common issues
Summary
Rewrites the ADVPERF01 performance-efficiency question page: adds a 'Capability intent' section, a 'Common issues to watch for' list, renames the question and best-practice links (BP05 now covers benchmarking managed vs self-hosted services), and updates the design-principles heading. The new common-issues list notes avoiding public-internet routing in favor of AWS PrivateLink for DSP/SSP traffic and Amazon Bedrock calls.
Security assessment
The change is a documentation restructure of a Well-Architected performance-efficiency page. The only security-adjacent content is the added guidance to use AWS PrivateLink rather than public internet endpoints for DSP/SSP and Amazon Bedrock traffic, which reduces network exposure, but it is framed as a latency/performance concern and references no specific vulnerability, CVE, or incident. No authentication, encryption, or credential changes are present.
Evidence
+ * Routing traffic between AWS-hosted DSP and SSP partners over the public internet instead of AWS PrivateLink, adding unnecessary latency and network exposure.
Diff
diff --git a/wellarchitected/latest/video-streaming-advertising-lens/advperf01.md b/wellarchitected/latest/video-streaming-advertising-lens/advperf01.md index ea9e804..4ecbc36 100644 --- a//wellarchitected/latest/video-streaming-advertising-lens/advperf01.md +++ b//wellarchitected/latest/video-streaming-advertising-lens/advperf01.md @@ -6,0 +7,2 @@ +Capability intentCommon issues to watch for + @@ -9 +11 @@ -The optimal solution for a particular workload varies, and solutions often combine multiple approaches. Well-Architected workloads use multiple solutions and allow different features to improve performance. +The optimal solution for a particular advertising workload varies, and solutions often combine multiple approaches. Well-Architected advertising workloads use multiple solutions and allow different features to improve performance. Architecture selection decisions directly affect the latency, burst capacity, and unit economics of programmatic advertising systems. Choosing the wrong architecture for geographic affinity, scaling strategy, compute, or AI/ML inference can result in missed bid deadlines, inefficient scaling, and avoidable cost penalties at scale. @@ -11 +13 @@ The optimal solution for a particular workload varies, and solutions often combi -ADVPERF01: How does your architecture account for advertising workload needs like low latency and bursty traffic? +ADVPERF01: How do you select and configure your architecture to meet the performance requirements of your advertising workload? @@ -12,0 +15,20 @@ ADVPERF01: How does your architecture account for advertising workload needs lik +[ADVPERF01-BP01 Design geographical affinity architecture with external entities (DSPs and SSPs)](./advperf01-bp01.html) +[ADVPERF01-BP02 Use appropriate scaling to handle burst traffic with cost considerations](./advperf01-bp02.html) +[ADVPERF01-BP03 Design for low latency with appropriate compute, storage, and network considerations](./advperf01-bp03.html) +[ADVPERF01-BP04 Evaluate AI/ML-based architecture for optimization](./advperf01-bp04.html) +[ADVPERF01-BP05 Benchmark managed and self-hosted services for latency-sensitive bid paths](./advperf01-bp05.html) + +## Capability intent + + * You have designed geographic affinity routing so that traffic from publishers and demand partners travels the shortest network path, keeping bid round-trip times within the real-time bidding window. + + * Your architecture scales elastically to absorb burst traffic from flash crowds and high-profile live events without dropped bid requests or latency degradation. + + * Every layer of your infrastructure (compute, storage, and network) is configured specifically for the low-latency, high-throughput demands of ad serving and bidding workloads. + + * AI/ML inference is served through a purpose-built, low-latency framework that meets programmatic auction deadlines and integrates with your observability stack. + + * You have made an evidence-based choice between self-managed open-source software and fully managed AWS services, validated through benchmarking under realistic advertising traffic. + + + @@ -14 +36 @@ ADVPERF01: How does your architecture account for advertising workload needs lik -Focus on the architecture design needs for your advertising workload, including networking, compute, storage, machine learning workloads, decision between self-managed and managed services. Address the low latency, burst traffic, and scaling needs of advertising workloads +## Common issues to watch for @@ -16 +38 @@ Focus on the architecture design needs for your advertising workload, including -###### Best practices + * Routing traffic between AWS-hosted DSP and SSP partners over the public internet instead of AWS PrivateLink, adding unnecessary latency and network exposure. @@ -18 +40 @@ Focus on the architecture design needs for your advertising workload, including - * [ADVPERF01-BP01 Design geographical affinity architecture with external entities (DSPs and SSPs) ](./advperf01-bp01.html) + * Deploying all compute in a single Region regardless of where bid requests originate, forcing cross-continental round trips. @@ -20 +42 @@ Focus on the architecture design needs for your advertising workload, including - * [ADVPERF01-BP02 Use appropriate scaling to handle burst traffic with cost considerations](./advperf01-bp02.html) + * Relying solely on reactive auto scaling without pre-warming load balancers, causing connection drops during sudden traffic bursts before scaling catches up. @@ -22 +44 @@ Focus on the architecture design needs for your advertising workload, including - * [ADVPERF01-BP03 Design for low latency with appropriate compute, storage, and network considerations ](./advperf01-bp03.html) + * Invoking [Amazon Bedrock](https://aws.amazon.com/bedrock/) over public internet endpoints instead of through AWS PrivateLink, adding per-call latency to the critical bid path. @@ -24 +46 @@ Focus on the architecture design needs for your advertising workload, including - * [ADVPERF01-BP04 Evaluate AI/ML-based architecture for optimization (like contextual advertising or scaling algorithms on event context) ](./advperf01-bp04.html) + * Defaulting to self-managed open-source or fully managed services without benchmarking either option under realistic advertising load. @@ -26 +48 @@ Focus on the architecture design needs for your advertising workload, including - * [ADVPERF01-BP05 Evaluate the choice of open source-based software (self-managed) against using a fully-managed service ](./advperf01-bp05.html) + * Hard-coding a specific x86 chipset (AMD or Intel), locking the workload into a single Spot pricing pool and missing cost savings when pricing shifts. @@ -37 +59 @@ To use the Amazon Web Services Documentation, Javascript must be enabled. Please -Performance efficiency +Performance efficiency design principles