SUMMARY
S3 has strong consistency for read after write, so PUT then GET is strongly consistent.
There is eventual consistency for PUTs and DELETEs, so if you upload or delete something, you still might see the old object because it does update instantly.
S3 supports parallel requests, which means you can scale your performance by the factor of your compute cluster, without making any customizations to your application.
Reserved Instance Marketplace is for selling Reserved EC2 instances which you don't need or for any other reason.
To protect your system from DDoS attack, you can use CloudFront for distributing content and use an ALB with Auto Scaling groups for your EC2 instances and also prevent direct Internet traffic to your database by deploying it to a new private subnet.
You can enable DynamoDB Auto Scaling to automatically scale your DynamoDB database with ease.
Kinesis Data Streams is a real-time data streaming service and you can also use a Lambda function to process the data.
S3 Single PUT limit is 5GB and the Multipart Upload Limit is 5TB.
CreationPolicy
is a CloudFormation resource attribute which waits upon a specified number of success signals or the timeout period is exceeded.cfn-init
reads from metadata and acts accordingly.cfn-signal
sees the changes to resources and acts accordingly.Redshift = OLAP, so stores data in columns.
S3 Glacier Provisioned Capacity enables you to always use Glacier expedited retrieval times.
To change the launch configuration for an ASG, you must create a launch configuration and then update your Auto Scaling group with it because you can't modify a launch configuration once you've applied it.
Active-active failover is used when you want ALL OF YOUR RESOURCES available all the time.
Active-passive failover is used when you want a primary resource available and when that isn't available, it should transfer to the secondary resource (standby).
File Gateway is for SMB and NFS and Volume Gateway is for block storage.
EC2 Instance store persists only when an instance reboots but if the instance does anything else, the data on the instance store is gone.
You can use AWS Backup for centralised management of all your backups and you can use it for RDS backups over 35 days and you use Backup plans to specify the details of the backups.
SWF can be used for managing tasks and jobs in AWS and for invoking one task with another and other things like that.
AWS limits how much vCPUs per region you can get but you can increase this limit with a support request to AWS.
OAI makes sure you can only access S3 objects through CloudFront and not directly to S3.
NACL rules are evaluated their rule number from lowest to highest.
You can grant cross-account access to S3 objects with IAM policies.
There are many different ways of SSE in S3:
Using KMS managed keys.
Using S3 managed keys.
Using customer provided keys.
Client Side Encryption in S3:
Using KMS managed CMK.
Using a client side master key (the most control because encryption + key is client side).
Last updated
Was this helpful?