Exam AWS Certified Solutions Architect - Professional topic 1 question 37 discussion

Answer is E Keyword is "or Availability Zone without any data loss" A: RAID 5 is not recommened by AWS. Also need replicate to another AZ B: Need synchronous replication to prevent any data loss (in case lost AZ) C: Need synchronous replication to prevent any data loss D: Need synchronous replication to prevent any data loss Also full option E is "Instantiate an i2.8xlarge instance in us-east-1a. Create a RAID 0 volume using the four 800GB SSD ephemeral disks provided with the Instance Configure synchronous block-level replication to an identically configured Instance in us-east-1b."

I've never seen any customer using RAID in the cloud, this is OLD. I'd rather fail the exam than provide such a solution.

E. Instantiate an i2.8xlarge instance in us-east-1a. Create a RAID 0 volume using the four 800GB SSD ephemeral disks provided with the instance. Configure synchronous, block-level replication to an identically configured instance in us-east-1b.

I would go with E which gives option for multi Az deployement

E is the answer. I used RAID instance storage in my work. I just don't feel comfortable with sync data cross zone. I thought about D too. So I did PoC creating the storage gateway. It is possible to satisffied the request with a newer instance type that can have large EBS IOPS but not for c3 familiy. I also checked with AWS support, there is no way for the user to use instance storage as a local disk

It has to be D. RAID 0 doesn't accept disk failures, so this is not an option. Answer A would work but doesn't prevent complete data loss because of the 15 minute backup window. D is the answer, there's no data loss, it has disk resilience through Storage Gateway and has the needed IO and disk space.

E is correct answer

Should it be C since the question ask for redundancy across AZ. There is no mentioned of redundancy for lost of region.

answer is D

How can 3TB provide 100,000iops?

I think AWS will remove this. Who would even consider synchronous replication when you need super fast disk arrays ( with synchronous repl, you wait until the write is completed on both local and remote which severely impact your super critical high perf disk array ) Asynchronous replication with writes being queued for a few minutes is much more acceptable

E is correct answer

Answer is E. Keyword is " survive loss of an AZ" which is possible only via E.

E, synchornous block-level replication to an Identical configure instance in other AZ

E makes sense. however, dont think that IOPS level could be met with a synch replication tbh. so keen to see what is the solution in details!

you do know Raid 0 has no resiliency right?

go with E