AMAZON EC2

RUNNING VIRTUAL SERVERS IN THE CLOUD

INTRODUCTION VIRTUAL MACHINES AMAZON MACHINE IMAGES INSTANCE TYPES ️ STORAGES
Elastic Compute Cloud 2 (EC2) is one of Amazon’s EC2 instances are virtual machines (VMs) in the cloud. AMIs are AWS-maintained configurations that are required Launching an instance requires you to specify an instance EC2 comes with a diverse set of possible storage options,
Infrastructure as a Service offerings for a broad range of Contrary to physical servers, virtual machine use software to to launch an instance. It contains the operating system (e.g. type. This determines the hardware capabilities of the host each with a unique combination of durability and
compute requirements, including more than 500 instances create an abstraction from their underlying hardware. This Amazon Linux 2), architecture (32/64-bit x86 or 64-bit that is used for your virtual machine.

performance.

with different processors, storages, operating systems and allows securely hosting multiple virtual machines (even from ARM), launch permissions, and storage for the root device.

networking capabilities. It’s one of the most mature services, different AWS customers) on the same physical hardware.

The instance types are grouped into instance families Regardless of your requirements, there will be a great fit
released back in 2006! Developers can also create shared images (Shared AMIs) General Purpose: a balance between compute, memory, available
An instance comes with strong performance and security that are made available for others to use. AWS can’t ensure and networking resources EC2 Instance Store: providing temporary (ephemeral)
guarantees, even when the underlying is shared, as AWS the integrity or security of the AMIs, so it’s within your own Compute Optimized: instances that offer high- storage for your instance, located on disks that are
INSTANCE LIFECYCLE dedicates some resources of the host computer such as CPU, responsibility. performance processors physically attached to the host computer. Use it for
Between the launch and the termination of your EC2
memory, and instance storage to the particular instance. Memory Optimized: for in-memory processing of large cached data, buffers, and temporary content. Data will be
instance, there more lifecycle states it transitions through
data sets lost if the underlying disk fails or the instance stops,
Launch pending Start CONNECTING TO YOUR INSTANCE Storage Optimized: best fit for workloads requiring high hibernates, or terminates
PURCHASES OPTIONS Key-pairs allow you to securely access your EC2 instances
performance reads and writes for large, locally saved, Elastic Block Storage (EBS): block storage volumes that
Reboot Stop
rebooting running
Hibernate
stopping terminated EC2 provides different purchase options that drastically from any location via SSH. If your launching a new instance,
data sets can be mounted to instances. Their persistence is
Terminate
vary in pricing. If you plan to use EC2 regularly, it’s a duty to you’ll be prompted to either create a new public & private
shutting down Accelerated Computing: instances that make use of independent from your instance’s lifetime. Use it for
understand the purchase options key set or use an existing one. When generating a new pair,
hardware accelerators and co-processors for fastest long-term storage of data that require low-latency
terminated
Terminate
On-Demand: the default option, billed by the second after AWS will store the public part on the instance (concisely at
processing of specific operations like graphics processing access, e.g. for databases
EBS-backing needed an instance is launched ~/.ssh/authorized_keys), while asking you to download the
or pattern matching, while also supporting the highest Elastic File System (EFS): a scalable file storage that can
Launch: creating and launching an instance from an AMI Reserved / Savings Plan: make a commitment to either private part. AWS won’t store the private part afterward, so
parallelism. scale based on workload requirements. Use it for any
into a VPC, subnet and availability zone. From creating an instance configuration (reserved) or usage in USD per if you lose it, you can’t recover it.

workload that can suddenly increase or decrease in

the machine to being available it will be in the pending hour (savings plan) for a dedicated time frame (1 to 3 You can also use AWS Systems Manager Session Manager
storage needs. It can also be used for shared volumes
state years) to significantly lower your bill in comparison to on- to connect via a browser-based shell or AWS CLI.

MONITORING Simple Storage Service (S3): data storage for

Stop & Start: with EBS-volumes you’re also able to start demand pricing There is a variety of services that helps you to monitor the
unstructured data, that are stored with no hierarchy and
and stop your instance. It keeps its instance ID and you Scheduled: reserve capacity based on a schedule for a As an alternative, you can solely rely on AWS IAM to availability and performance of your instances.

are only accessed by a unique object identifier.

won’t be billed for the instance while it’s stopped, but year. This is currently not available at AWS and not connect to your instances, not needing to manage any key-
you’ll still pay for the storage of any EBS volume planned to be reactivated pairs. This is achieved with EC2 Instance Connect. The simplest tools are system status checks and instance
Hibernate: If supported by the OS, the instance can Spot: use spare EC2 capacity for pricing based on offer & status checks. System status checks are provided and taken AUTO-SCALING ️
saved its RAM contents to the EBS root volume. This pre- demand, much lower than on-demand. Keep in mind that care of by AWS and detecting problems with the software or You’re able to organize your instances into logical groups
warmed state takes much quicker to initalize if it’s again your workloads can be interrupted if there is not enough
AWS MARKETPLACE physical hardware, e.g. network connectivity or power loss. that can be used to scale based on policies, including
While many AWS-provided AMIs come with additional
needed in production and isn’t billed spot capacity available Instance status checks are there to detect configuration Target Tracking: specify target values for a metric, e.g.
software, especially for software development purposes, it
Reboot: restarting your operating system. Your instance Dedicated Instances & Hosts: launch instances onto issues that need to be addressed by yourself, including average CPU or memory consumption
most likely doesn’t include all the software you need.

will keep private and public IP addresses physical servers that are isolated at hardware level. misconfigured networks or exhausted memory.

Step Scaling: define one or several thresholds for metrics

Retire: if underlying hardware experiences irreparable Dedicated instances might share hardware with other that will then trigger a scale-in or out event
As described earlier, shared AMIs come to tackle this issue.
failure, AWS will schedule your instance for termination instances in the same AWS account, while hosts are not CloudWatch and EventBridge help you gain more insights Scaling based on SQS: if your instances workloads are
Additionally, there are shared images that are also paid and
Terminate: deleting your instance if no longer needed On-Demand Capacity Reservations: pre-plan EC2 into your instances and also react to incidents with received via SQS, you can scale based on the queue’s
therefore increase the price for your instance types. When
Recover: instances can automatically recover either on capacity for your instances in a availability zone. Those automated routines. EC2 automatically sends metrics to depth. This helps you add more workers on a fast-
selecting a paid AMI from AWS Marketplace you’ll be
status check or hardware failures. reservations can be created at any time (no yearly CloudWatch, including CPU utilization as well as network growing queue and remove workers if the queue’s state
informed about the additional usage fees.
commitments needed) and are independent from the and disk usage. is less busy
discounts of saving plans or reserved instances. Scheduled Scaling: scale based on time and date if you’re
aware of traffic or workload patterns.