DevOps일반2025년 DevOps 공부 아이템

2025 DevOps 연구(역량 강화) 계획

1. GitOps / IaC(Infrastructure as a Code)

IaC - Terraform, Ansible 등 git의 코드로 동작하는 DevOps 전반
GitOps – the practice of using Git repositories to manage infrastructure configurations and operations – has been gaining popularity and is expected to become a standard practice.

2. Multi-Cloud DevOps / SA와 인프라 업무

AWS, Azure 공부

Redis PaaS

VM에 docker container로 올리는 대신 다음 PaaS로 변경 (가격이 비슷함)

  • AWS Elasticache for Redis / Memcached
  • Azure Cache for Redis

SK렌터카

Azure / DevOps

Rolling update시 404 error 해결 후보들

금천미트에서 해결하지 않은 것

1. Blue/Green Canary deployments

2. Serve static assets from an external CDN (or external storage)

3. Layer 7 LB (Load Balancer)
  • AWS ALB, Azure Application Gateway와 session stickiness 기능

3. Serverless Architecture

AWS Lambda, serverless Aurora 등, 기존 infrastructure overhead를 줄일 수 있다.

4. Python

for automation such as AWS Lambda - gitOps에 여러 부분을 자동화

5. Kubernetes Maturity

Container orchestration에서 쿠버네티스는 계속 표준이 될 것이고 기능이 강화될 것 (아이스크림미디어는 1.29, 현재 1.32 나오는 중)

5-1. Telemetry

Prometheus, Grafana, Kiali, Jaeger
AWS CloudWatch를 통한 모니터링 서비스

5-2. Service Mesh

Istio 기능 고도화

5-3. Adaptive Scaling

Dynamic resource scaling based on predictive analytics will become a core capability.

6. Network

VPN 구성, Load Balancer, Firewall, Azure NIC, Azure NSG, Azure PrivateLink
AWS와 Azure의 Kubernetes의 Network Plugin / Network Policy
CNI 종류 Calico / Cilium / AWS VPC CNI / Azure CNI → which supports virtual node?

7. AI Cloud Services / GenAI adoption in AIOps

AWS Bedrock - foundation model과 multi-modal integration
AWS SageMaker - 학습에 필요한 리소스

8. DevSecOps / Cybersecurity

오픈 이전에 모의해킹이 아닌 개발 단계에서 꾸준한 보안 체크 필요

https://youtu.be/kqaMIFEz15s

  1. (unauthorized) Shadow AI - pull from a cloud instance and go away (data leakage)
  2. Deepfake - $25million $35million
  3. Exploits/Malware - GenAI can write malware
    Amazon has reported a significant increase in potential cyber threats in 2024, detecting nearly 1 billion possible incidents daily, up from 100 million earlier this year. This surge is partly attributed to the growing use of artificial intelligence by attackers. https://www.wsj.com/articles/the-ai-effect-amazon-sees-nearly-1-billion-cyber-threats-a-day-15434edd?utm_source=chatgpt.com
    Similarly, British telecom company BT has observed a substantial rise in cyber-attack signals, detecting 2,000 potential attacks per second on its network. In July 2024, BT noted a 1,200% increase in malicious scanning activities compared to the same month the previous year.
    https://www.reuters.com/technology/cybersecurity/bt-spots-2000-potential-attacks-its-network-second-2024-09-12/?utm_source=chatgpt.com
  4. Prompt injection attack → OWASP에서 LLM을 상대로한 1위 공격 방법이 될 것이다. 2025년 보안의 최대 이슈.
  5. Increase of Attack Surface
  6. AI phishing
  7. (1-6까지는 Risk, 6은 강화) AI 추천 response for cyberattacks → 추천 대응 방안을 내놓으면 전문가가 택해서 실행

9. Edge computing / Distributed Cloud

Unlike a centralized environment, Edge Computing distributes processing and data collection across various points at the network’s edge, reducing latency and, consequently, increasing efficiency.
Distributed Cloud(중앙+edge) > Edge > CDN
CDN에서 확장 개념 - 동적 real-time processing. ioT 환경에서 우선 도입 중.