AWS Summit Seoul 2024 — A Backend Engineer's Field Notes (Generative AI · CMP · Middleware)

📅 Notes from attending AWS Summit Seoul on May 17, 2024. Backend engineer's view — what booths people had, what the products were, what the actual competitive landscape looks like.

Figure 1. AWS Summit name badge (name redacted).

Why I went

Less for the sessions, more to walk the booths and read the market. Who's bringing what, and who's competing with whom in adjacent niches.

Figure 2. AWS photo zone.

Figure 3. Exhibitor list — heavy on cloud SI / MSP: Megazone, SK C&C, KT DS, NHN Cloud, LG CNS, …

What I saw — three buckets

1. Generative-AI serving — "everyone's doing the same four things"

The application patterns at every booth converged on the same shortlist: - Chatbots — internal knowledge Q&A, customer service - Email automation — summarize, draft, classify - Personal copilots - Image generation

Figure 4. "Past-life experience" generative-AI demo — every booth had some fun, interactive hook.

The fight is who can ship it faster, prettier, and more accurate. The core ideas are nearly identical.

2. CMP — Cloud Management Platform

CMP (Cloud Management Platform) = a single pane of glass for managing multiple public clouds (AWS · Azure · GCP · OCI · …). Cost · resources · security · policy, all consolidated.

Key players: - Megazone Cloud — Korea's top AWS MSP, with its own CMP - SK Inc. C&C - KT DS

As more enterprises use multi-cloud, "pick the best service from each public cloud" becomes feasible. Avoids single-cloud lock-in and gives leverage on price + region.

3. Middleware — API gateway & service mesh

Classic enterprise-backend territory, still very active: - Routing (send requests to the right backend) - Authentication / authorization (OAuth · JWT) - Rate limiting (DDoS protection · billing) - Circuit breaker (stop failure propagation)

Solutions named: - Nginx — still the standard - KT DS Vesta — in-house API Gateway product

One-line takeaway

"All software ideas circle back to each other."

The features converge; UX and performance are where the differentiation actually happens. That's a meaningful insight for a backend engineer too — it's not just the core algorithm; developer experience, ops convenience, real-world performance are decisive.

References

• AWS Summit Seoul: https://aws.amazon.com/events/summits/seoul/
• Companion review post:
  
  KSME Conference Review — Neural-Net PM2.5 Prediction (Gold Award, KSME Fluid Mechanics Competition, 2019)

  An undergraduate thesis from 2019 — using a neural network (ANN) to predict PM2.5 concentrations 6/12 hours ahead. 94,225 observations from weather + air-quality data, AIC·VIF variable selection, Window Learning. Plus the time-lag pitfall discovered two years later.

  taystudios.com/blog

📦 Migrated from my own Korean blog (my own writing). Original: taehyuklee.tistory.com/11