Accounts, Billing, and Support Flashcards
Which of the following is not one of the three AWS Pricing Fundamentals:
Storage
Compute
Inbound Data Transfer
Outbound Data Transfer
Inbound Data Transfer
AWS Pricing Fundamentals:
- Compute - amount of resources (CPU, RAM) and duration of use
- Storage - quantity of data stored (allocated)
- —>S3 = amount of data you uploaded
- —>EBS = the size of the volume you create - Outbound Data Transfer - quantity of date going outbound from services
Pricing Models:
Pay-as you go
Save when you reserve
Pay less by using more
Pay-as-you-go:
- -Easily adapt to changing business needs, not forecasts
- -Right-sizing
- -Improved responsiveness to change
- -Reduce risk over over-positioning of missing capacity
Save when you reserve:
- -You reserve capacity for a reduced rate (eg. RDS and EC2)
- -Save up to 75% compared to on-demand (pay-as-you-go) rates
- -The more you pay upfront the greater the discount
Pay less by using more:
- -Pay less using volume-based discounts
- -Tiered pricing means the more you use the lower the unit pricing
- Ex. Unit price (per GB of storage) could be lower is you are using more
EC2 Pricing Options:
On-Demand Reserved Scheduled Reserved Instances Spot Instances Dedicated Instances Dedicated Hosts Savings Plans
Covered in more detail on separate flashcards
EC2 On-Demand Pricing:
EC2 On-Demand Pricing:
- Standard rate - no discount
- No commitments
- Not intended for long running workloads b/c most expensive option
Best for:
- ->Dev/test
- ->Short-term, or unpredictable workloads
- ->Small projects
EC2 Reserved Pricing:
EC2 Reserved Pricing
- 1 OR 3 year commitment
- Up to 75% discount
Best for:
- ->Steady-state, predictable workloads
- —->Ex. Line of Business applications that are predictable and run year after year so need to secure the best pricing model
- ->Reserved capacity w/ reserved instances
EC2 Scheduled Reserved Instances Pricing:
EC2 Scheduled Reserved Instances Pricing
- Match capacity reservation to recurring schedule
- Minimum 1200 hours per year
-Ex: Reporting application that runs 6 hours a day/4 days a week = 1248 hours per year
So, you reserve 1248 hours per year and get a reserved discount
You can ONLY use the service during the agreed upon times
AWS is phasing out this option
EC2 Spot Instances Pricing:
EC2 Spot Instances Pricing
- You bid for unused capacity
- 90% discount
- Negative - Instances can be terminated at any time and you get very little notice
- —>2 minute warning if AWS needs to reclaim capacity
- —>warning is available via instance metadata and CloudWatch Events
Positive - Very cost effective solution to run a large processing workload
Spot Block
- You can run EC2 instances uninterrupted for 1 - 6 hours
- Specify the block duration, number of instances and your bid price
- Pricing is 30% - 45% less than On-Demand (not as big of a discount as Spot but if you need uninterrupted timeframe this is better option)
BEST FOR:
- Workloads w/ flexible start and end times
- Data is saved elsewhere and application logic is able to recover from any terminated instances
Spot Instance
VS
Spot Fleet
VS
EC2 FleetSpot Instance
-One or more EC2 instances
Spot Fleet
-AWS launches and maintains the number of Spot/On-Demand instances to meet specified target capacity
EC2 Fleet
- AWS launches and maintains a specified number of Spot/On-Demand/Reserved instances in a single API call
- Can define separate On Demand/Spot capacity targets, bids, instance types, and AZs
EC2 Dedicated Instances Pricing:
EC2 Dedicated Instances Pricing
-Physical isolation at the Host Hardware level from instances belonging to other AWS customers
—->(as opposed to sharing the infrastructure on which AWS is run with other customers)
—->Dedicated Instances can share hardware w/ other instances from YOUR account but not from other AWS customers
- Pay per INSTANCE
- Expensive
EC2 Dedicated Hosts Pricing:
EC2 Dedicated Hosts Pricing
- Physical Server dedicated for your use
- Pay per HOST
- Can add capacity using an allocation request
- MOST Expensive of the options
BEST FOR:
- Workloads w/ server-bound software licenses
- —>Some licensing is bound to CPUs in physical Host where you instance is run
- —>Visibility of sockets for licensing, cores, host ID - Host affinity
- —>Control which Host a particular instance is running on
- —>May want to make sure 2 instances are running on diff hosts (targeted instance placement)
EC2 Savings Plans Pricing:
EC2 Savings Plans Pricing
- Hourly commitment to a consistent amount of usage
- 1 OR 3 yr commitment
- COMPUTE Savings Plan
- —>Hourly commitment to usage of EC2, Fargate and Lamda
- —>Usage in any Region or family
- —>Any size, tenancy, and OS - EC2 Savings Plan
- —>Hourly commitment to usage of EC2 ONLY
- —>Usage for Ec2 only w/in a selected Region and Instance Family
- —>Any size, tenancy, and OS
Amazon EC2 Billing:
Amazon EC2 Billing
- Billed per second; minimum of 1 minute
- —>Amazon Linux and Ubuntu in On-Demand, Reserved, and Spot Forms
- —>EBS Volumes attached to EC2 instances are also billed per second; minimum of 1 minute
Billed per hour; minimum of 1 hour
- —>Windows
- —>Commercial Linux distributions ie:
- ——->Red Hat Enterprise Linux (EL)
- ——->SUSE Enterprise Server (ES)
Amazon EC2 Reserved Instances (RIs) Billing:
Amazon EC2 Reserved Instances (RIs)
Payment can be:
- ->All Upfront - BEST discount but you have to pay full amount up front (not great for your cash flow)
- ->Partial Upfront - Partial fee upfront and then a monthly payment
- ->No Upfront - No partial fee upfront, just monthly payment (better for your cash flow)
Two Types of Reserved Instances: 1. Standard RI -->Can change: ---->Availability Zone ---->Instance size (if Linux) ---->Networking type USE MODIFY_ RESERVED_INSTANCES API action to perform changes
- Convertible RI
–>Can change:
—->Same as Standard PLUS:
—->Family
—->Operating System (OS)
—->Tenancy
USE EXCHANGE_RESERVED_INSTANCES API action to perform changes
When the attributes of an instance you launched in your account match the attributes of a reserved instance that is when any discounts are applied
Attributes include:
–>Type of instance you are running
–>Operating System you are running it on
–>Availability Zone
–>Region
–>Tenancy:
—->Default
—->Dedicated - running on dedicated hardware
When an Availability Zone is specified you can reserve capacity in that AZ
- —>CANNOT reserve capacity in Region
- —>RI discount applies to all Availability Zones but not Regions
Amazon S3 Pricing:
Amazon S3 Pricing
Charged for:
- Storage class that you are using
- —>e.g. Standard or IA
- Storage Quantity
- —>data volume stored in your buckets on a per GB basis
- Number of Requests
- —>number and type of requests
Lifecycle transitions requests
—->requests to move data between storage classes
- Data transfer
- —>data transfer out of an S3 region is charged
Amazon S3 Glacier Pricing:
Amazon S3 Glacier Pricing
Three options for access to archives:
- Expedited - 1-5 minutes data access time (not available for Glacier Deep Archive)
- Standard - 3-5 hours data access time
- Bulk - 5-12 hours data access time
