HA architecture

Question 1

In discussions about Cloud services the words ‘Availability’, ‘Durability’, ‘Reliability’ and ‘Resiliency’ are often used. Which term is used to refer to the likelihood that a resource will work as designed?

Answer 1

Reliability

Each word has a specific meaning and your ability to select a correct answer may depend on understanding the difference. Reliability is closely related to Availability, however a system can be ‘Available’ but not be working properly. Reliability is the probability that a system will work as designed. This term is not used much in AWS, but is still worth understanding.

Question 2

Which service works in conjunction with EC2 Autoscaling in order to provide predictive scaling based on daily and weekly trends?

Answer 2

AWS Autoscaling

Correct Answer: B
Why is this correct?
EC2 Autoscaling works in conjunction with the AWS Autoscaling service to provide a predictive ability to your autoscaling groups.

Question 3

Placement Groups can either be of the type ‘Cluster’, ‘Spread’, or ‘Partition’. Choose options from below which are only specific to Spread Placement Groups.

Answer 3

A spread placement group is a group of instances that are each placed on distinct underlying hardware

Question 4

Following an unplanned outage, you have been called into a planning meeting. You are asked what can be done to reduce the risk of a single bad deployment taking the whole site down. (The selected options do not necessarily need to work together)

Answer 4

Use several Target groups or auto scaling groups under each Load Balancers.

Use a Classic Load Balancer to spread the load over several availability zones.

Use Route 53 with health checks to distribute load across multiple ELBs.

Use multiple autoscaling groups and boundaries for a staged or ‘canary’ deployment process.

Question 5

components of EC2 Autoscaling?

Answer 5

Scaling Options
Groups
Configuration Templates

Question 6

A product manager walks into your office and advises that the simple single node MySQL RDS instance that has been used for a pilot needs to be upgraded for production. She also advises that they may need to alter the size of the instance once they see how many people use the system during peak periods. The key concern is that there can not be any outages of more than a few seconds during the go-live period. Which of the following might you recommend,

Answer 6

Consider replacing it with Aurora before go live.

Convert the RDS instance to a multi-AZ implementation.

Question 7

You work for a manufacturing company that operate a hybrid infrastructure with systems located both in a local data center and in AWS, connected via AWS Direct Connect. Currently, all on-premise servers are backed up to a local NAS, but your CTO wants you to decide on the best way to store copies of these backups in AWS. He has asked you to propose a solution which will provide access to the files within milliseconds should they be needed, but at the same time minimizes cost. As these files will be copies of backups stored on-premise, availability is not as critical as durability. Choose the best option from the following which meets the brief.

Answer 7

Copy the files from the NAS to an S3 bucket with the One Zone-IA class

S3 OneZone-IA provides on-line access to files, while offering the same 11 9’s of durability as all other storage classes. The trade-off is in the availability - 99.5% as opposed to 99.9%-99.99%. However in this brief as cost is more important than availability, S3 OneZone-IA is the logical choice . RRS is deprecated and new uses are strongly discouraged by AWS.