Amazon Web Services
lastmod 2026-04-23created 2020-03-2066 edits
Amazon Web Services is the cloud computing division of Amazon.com. Amazon Web Services provides online services to other websites and client-side applications.
Categories
- Amazon Elastic Compute Cloud (Amazon EC2; Elastic Compute Cloud EC2) - Virtual Machine
- EC2:GettingStarted - How to get started with EC2.
- EC2:InstanceType
- EC2 Computing Unit (ECU)
- Easily set up a single virtual server by choosing one of the given resource options (Ubuntu, Node, Lamp stack, Nginx, WordPress, Django.. etc)
- Creates a virtual waiting room and buffers user requests when heavy traffic occurs
Databases
| Database type | Use cases | AWS services |
| Relational | Traditional applications, enterprise resource planning (ERP), customer relationship management (CRM), e-commerce | Amazon Aurora, Amazon RDS, Amazon Redshift |
| Key-value | High-traffic web applications, e-commerce systems, gaming applications | |
| In-memory | Caching, session management, gaming leaderboards, geospatial applications | Amazon ElastiCache, Amazon MemoryDB for Redis |
| Document | Content management, catalogs, user profiles | Amazon DocumentDB (MongoDB compatible) |
| Wide column | Large-scale industrial apps for equipment management, fleet management, and route optimization | Amazon Keyspaces |
| Graph | Fraud detection, social networking, recommendation engines | Amazon Neptune |
| Time series | Internet of Things (IoT) applications, DevOps, industrial telemetry | Amazon Timestream |
| Ledger | Systems of record, supply chain, registrations, banking transactions | Amazon Ledger Database Services (QLDB) |
Networking
- AWS networking and content delivery
- Amazon VPC - Customize and control your networking environment with Amazon Virtual Private Cloud (VPC)
- AWS Transit Gateway - Simplify your network by connecting VPCs and on-premises networks to a single gateway
- AWS PrivateLink - Establish private connectivity between VPCs and AWS or on-premises services
Application networking
- AWS AppMesh - Connect containers and microservices with application-level networking
- AWS API Gateway - Create, maintain, and secure APIs at scale
- AWS Cloud Map - Discover and access up-to-date service resources
Edge networking
- Amazon CloudFront - Deliver data, videos, applications, and APIs with low latency and high transfer speeds
- Amazon Route 53 - Improve end users' use of internet applications with a low-cost managed Domain Name System (DNS)
- AWS Global Accelerator - Optimize user traffic to applications
Hybrid connectivity
- AWS Direct Connect - Establish a private, dedicated AWS connection to data centers, offices, or colocation environments
- AWS Site-to-Site VPN - Create encrypted network connections to Amazon VPC or AWS Transit Gateway
- AWS Client VPN - Connect your remote workforce to AWS or on-premises through a virtual private network (VPN)
- AWS Cloud WAN - Easily build, manage, and monitor a global wide area network
Network security
- AWS Shield - Protect AWS applications from distributed denial-of-service (DDoS) attacks
- AWS WAF - Protect web applications from common web vulnerabilities
- AWS Network Firewall - Deploy network security across Amazon VPCs
- AWS Firewall Manager - Centrally configure and manage firewall rules
Pricing
- Amazon EC2 Pricing – Amazon Web Services
- [Recommended] AWS Pricing Calculator <- Try running the numbers with the pricing calculator.
- DLAMI pricing - Deep Learning AMI
- Amazon EC2 Spot Instances Pricing – Amazon Web Services
- Instance purchasing options - Amazon Elastic Compute Cloud
- How should you choose an Amazon EC2 instance product and pricing plan?
Amazon EC2 provides the following purchasing options so that you can optimize costs based on your requirements.
- On-Demand Instances - Pay, by the second, for the instances that you launch.
- Savings Plans – Reduce your Amazon EC2 costs by committing to a consistent amount of usage, in USD per hour, for a term of 1 or 3 years.
- Reserved Instances – Reduce your Amazon EC2 costs by committing to a consistent instance configuration, including instance type and Region, for a term of 1 or 3 years.
- Spot Instances – Request unused EC2 instances to reduce your Amazon EC2 costs significantly. - However, prices change from moment to moment.
- Dedicated Hosts - Pay for a physical host that is fully dedicated to running your instances, and bring your existing per-socket, per-core, or per-VM software licenses to reduce costs.
- Dedicated Instances - Pay, by the hour, for instances that run on single-tenant hardware.
- Capacity Reservations – Reserve capacity for your EC2 instances in a specific Availability Zone for any duration.
Cost reduction references
GPU instances
- How to use Amazon EC2 Spot Instances to train deep learning models on GPUs
- Amazon EC2 P3 Instances - Accelerate machine learning and high performance computing applications with powerful GPUs
- AWS now offers NVIDIA Quadro Virtual Workstation for EC2 G4 instances at no additional cost (January 2020 article)
- Recommended GPU instances
- Building a GPU-based deep learning training environment with AWS
GPU instances are recommended for most deep learning purposes. Training new models is faster on GPU instances than on CPU instances. You can scale sub-linearly when you have multi-GPU instances or when you use distributed training across many instances with GPUs. To set up distributed training, refer to the section. Choose Distributed training.
The following instance types support the DLAMI. For more information on GPU instance type options and their uses, choose EC2 instance types and choose Accelerated Computing. Note
| INFORMATION |
| The size of your model is a factor in choosing an instance. If your model exceeds an instance's available RAM, choose a different instance type with enough memory for your application. |
- Amazon EC2 P3 instances provide up to 8 NVIDIA Tesla V100 GPUs
- Amazon EC2 G3 instances provide up to 4 NVIDIA Tesla M60 GPUs
- Amazon EC2 G4 instances have up to 4 NVIDIA T4 GPUs
- Amazon EC2 P4 instances provide up to 8 NVIDIA Tesla A100 GPUs
DLAMI instances provide tools to monitor and optimize GPU processes. For more information on monitoring GPU processes, refer to the section. Choose GPU monitoring and optimization.
- G3, G4: For graphics-intensive applications such as 3D visualization.
- P3, P4: For machine learning training.
Email about the pricing plan research
크게 P3, P4, G3, G4 라인업 있음.
* G3, G4 는 3D 그래픽 전용.
* P3, P4 는 머신러닝 전용.
따라서, P3 또는 P4 사용해야함.
* P3 -> NVIDIA Tesla V100 GPU 사용
* P4 -> NVIDIA Tesla A100 GPU 사용
1개월 Demo 전제로, "P3 인스턴스" 요금제를 조사.
[Amazon EC2 P3 인스턴스 제품 세부 정보]
- 참조 페이지: https://aws.amazon.com/ko/ec2/instance-types/p3/
"p3.2xlarge" 인스턴스가 적합.
P3.2xlarge 으로 온 디맨드 사용시:
- 1개월, 3.06USD ×24시간×31일×1200원 환율 -> 2,731,968 원 사용.
- 1년, 2731968원×12개월 -> 32,783,616 원 사용.
- 참고: 사용 시간별로 요금이 차감됨.
P3.2xlarge 으로 1년 약정 사용시:
- 1개월, 1.99USD×24시간×31일×1200원 환율 -> 1,776,672 원 사용.
- 1년, 1776672원 ×12개월 -> 21,320,064 원 사용.
- 참고: 강제 1년 계약.
결론: p3.2xlarge, 온 디맨드 계약이 가장 적절해 보임.
이상입니다.
Regional Endpoints
| Name | Code |
| US East (Ohio) | us-east-2 |
| US East (N. Virginia) | us-east-1 |
| US West (N. California) | us-west-1 |
| US West (Oregon) | us-west-2 |
| Asia Pacific (Hong Kong) | ap-east-1 |
| Asia Pacific (Mumbai) | ap-south-1 |
| Asia Pacific (Osaka-Local) | ap-northeast-3 |
| Asia Pacific (Seoul) | ap-northeast-2 |
| Asia Pacific (Singapore) | ap-southeast-1 |
| Asia Pacific (Sydney) | ap-southeast-2 |
| Asia Pacific (Tokyo) | ap-northeast-1 |
| Canada (Central) | ca-central-1 |
| Europe (Frankfurt) | eu-central-1 |
| Europe (Ireland) | eu-west-1 |
| Europe (London) | eu-west-2 |
| Europe (Paris) | eu-west-3 |
| Europe (Stockholm) | eu-north-1 |
| Middle East (Bahrain) | me-south-1 |
| South America (São Paulo) | sa-east-1 |
Adding sub-users to a contracted instance
Go to My Security Credentials from the user menu at the top to enter Identity and Access Management (IAM).
You can add users here.
What is a CPU credit?
See AWS:Credit.
Tenancy
- Shared tenancy (multi-tenancy)
- Multiple AWS accounts share the same physical hardware.
- Other users' instances may run on the physical hardware.
- Dedicated Instance tenancy
- Instances run on single-tenant hardware.
- Other users' instances cannot run on the physical hardware. However, other instances of the same user can.
- Dedicated Host tenancy
- Instances run on a physical server with EC2 instance capacity fully dedicated to your use.
- No other instances (including both other users' and the same user's) can run on the physical hardware.
See also
- OpenFLOW - A tool for creating AWS-style Korean: 아이소메틱 infrastructure diagrams
- aws-doctor - AWS cost diagnostics, idle resource detection, and cloud spend optimization
- Floci - A free, open-source local AWS emulator
- MiniStack - A free, open-source local AWS emulator
- moto - Mock AWS Services
- LocalStack
Favorite site
- Wikipedia (en) Amazon Web Services
- Techniques for working with Amazon Web Services - Table of contents 1
- Amazon Web Services Korea Blog - Optimization strategies for AWS-based game backends as traffic decreases
- 9 basic Amazon web server (Amazon Web Service, AWS) services that may help developers
- (AWS/Amazon Web Services) Memories of absurd charges. Bill shock
- How to elegantly manage developers' AWS permissions with GitOps | Hyperconnect Tech Blog
- AWS in 2025: Things you think you know that are now wrong | GeekNews
References
-
Pyrasis-com.tar.gz ↩
- Aiobotocore
- Amazon
- Amazon CloudFront
- Amazon Elastic Compute Cloud
- Amazon Elastic Container Registry
- Amazon Elastic Kubernetes Service
- Amazon Resource Name
- Amazon Route 53
- Amazon Simple Queue Service
- Amazon Simple Storage Service
- AWS Amplify
- AWS CodeCommit
- AWS Command Line Interface
- AWS Fargate
- AWS Identity and Access Management
- AWS Lambda
- Aws-doctor
- AWS:Amplify:Domain
- AWS:CLI:Amplify
- AWS:Credit
- Boto3
- Cloud Computing
- Data Version Control
- EC2 Computing Unit
- EC2:GettingStarted
- EC2:InstanceType
- FastAPI
- Floci
- Infrastructure as a Service
- Karpenter
- Kubernetes:Basic
- Kumo
- LocalStack
- Lumberyard
- MiniStack
- Moto
- OpenFLOW
- Pgactive
- PurchaseRequisition:AWS
- Python:Libraries
- Wowza Media Systems