Amazon Compute Services
- AWS Elastic Compute Cloud (EC2) Overview
- It is a virtual machine in the cloud
- Uses pay as you go model
- Integrates with storage, networking, and security
- Implementation is very fast
- Some of the supported Operating Systems are:
- Windows 2003 R2 – 2016
- Amazon Linux
- Debian
- SUSE
- CentOS
- Red Hat Enterprise Linux
- Ubuntu
- EC2 Benefits
- Time to market (implementing any idea)
- Scalability
- Reliability
- Security
- Full control
- Services integration (other integrations managed by AWS like RDS)
- Cost efficiency
- EC2 instance types
- General Purpose
- T2, M5, M4, and M3
- Provides a balance of memory and network resources
- Use for less busy sites (websites, database servers, files servers)
- T2 provides burst performance
- Credits accrue during idle times
- Credits used to burst performance
- M5, M4, and M3 have no burst option
- Good for development and staging
- Compute Optimized
- C5, C4, and C3
- Useful for CPU-intensive applications
- Media coding
- Intensive batch jobs
- Many concurrent users
- Gaming servers
- Memory Optimized
- X1e, X1, R4, and R3
- Useful for high memory requirements
- Processing large data sets
- In-memory databases
- Big data processing
- Storage Optimized
- H1, I3, and D3
- Useful for high sequential read/writes to local storage
- Relational databases
- Data warehousing
- Image storage and processing
- General Purpose
Pro Tip: Memory and Storage Optimized types can be overlap because you can use memory-optimized and then implement EBS volume which gives storage optimization.
- Advanced Computing
- P3, P2, G3, and F1
- Useful for specialty hardware compute requirements
- Graphics Processing Unit (GPU)
- Field-Programmable Gate Array (FPGA)
Pro Tip: You don’t need to memorize what all the components do. Only find your particular components according to your need.
Note: New instance types come over time. So be sure to see documentations
- EC2 Pricing Categories
- On-demand
- Charged for usage time at a flat rate
- Billed in 60-second increments rounded up
- Reserved
- Hours (how much I am going to need for the next year, 2 years, or 3 years)
- You are reserving usage time in advance
- Reserved usage minimum 1 year
- Spot
- Bid on unused compute time
- Up to a 90% discount over on-demand category
- On-demand
- Shared tenancy
- Multiple customers share the time and space on the physical machine
- This is the default instance behavior
- Pros
- Reduced costs
- Simpler deployment
- Cons
- Lower performance
- Less control on a physical level
- Not suited for compliance
- Dedicated hosts
- Physical machines managed by AWS
- Used by only one AWS customer (You / Your organization)
- Must be explicitly configured
- Not available in a free tier
- Pros
- More accurate licensing management
- Because some licensing applied on the hosts level to run virtual machines/services (but this model is outdating)
- More detailed reporting in licensing
- Compliance management
- Determine host placement during instance restarts
- More accurate licensing management
- Cons
- Costs more
- More management on a physical level
Pro tip: Organizations can bring their license to cut the cost either on dedicated hosts or virtual machines. For example, they can use Windows machines which are less expensive and use their license.
- Dedicated instances
- It does run on a physical machine same as shared tenancy
- Only your machine runs on that host
- On a restart, may be moved to a different host
- Used by one customer
- Must be explicitly configured
- Not available in a free tier
- Pros
- Runs on hardware dedicated to the customer
- Provides performance advantage of a dedicated host
- Cons
- Less accurate licensing management
- It does run on a physical machine same as shared tenancy
- Amazon Machine Image (AMI)
- Blueprint with server configuration details
- Similar to localized imaging solutions
- The term “Instance” indicates the use of the AMI
- All instances are created from an AMI
- Sources are:
- Amazon (free): built by Amazon
- Marketplace (free/$): built by vendors (mostly)
- Community (free): built by anyone
- Who can launch an instance of an AMI?
- Public: Anyone can launch
- Explicit: Specified person/group can launch
- Implicit: Only the owner can launch
- Methods to create instances
- Use existing AWS AMIs
- Customize existing AMIs
- Create from scratch
- Use from other public sources
- Use this method with caution!
- Concepts of how virtualization works
- Hardware Virtual Machine (HVM) AMIs
- AMIs fully virtualizes the hardware
- Required hardware-assisted virtualization
- Same like CPU has special capabilities to work efficiently
- Paravirtual (PV) AMIs
- Run on hosts without specific support for virtualization
- Doesn’t perform as well as HVM AMIs
- Hardware Virtual Machine (HVM) AMIs
Pro Tip: You can change the instance type after creating the instance but be sure to stop before changing the instance type.
- Elastic Container Service (ECS)
- It is a way to implement docker container in the AWS
- ECS features
- No virtual machine builds required
- Uses Amazon Fargate to automatically build environments
- You don’t have to find Amazon Fargate because if you launch ECS instance, it will use Fargate in the backend
- You can use EC2 instances for docker for more control
- Elastic Beanstalk environment
- Deploy, monitor, and scale quickly and easily your application
- It pre-configured the environment for you to run your application
- It supports multiple services like Java, .NET, PHP, Node.js, Phyton, Go, and Docker on familiar servers such as Apache, Ngnix, Passenger, and IIS
- After configuration, you can see this instance in EC2 dashboard
Note: The work beanstalk came from where Jack plants the bean in the ground and it becomes a large stalk and grows like a giant plant. Same as with very little effort you build something really big.
0 Comments