Which of the following designs will meet these objectives’?

You have been asked to design the storage layer for an application. The application requires disk performance
of at least 100,000 IOPS in addition, the storage layer must be able to survive the loss of an individual disk. EC2
instance, or Availability Zone without any data loss. The volume you provide must have a capacity of at least 3
TB.Which of the following designs will meet these objectives’?

You have been asked to design the storage layer for an application. The application requires disk performance
of at least 100,000 IOPS in addition, the storage layer must be able to survive the loss of an individual disk. EC2
instance, or Availability Zone without any data loss. The volume you provide must have a capacity of at least 3
TB.Which of the following designs will meet these objectives’?

A.
Instantiate an 12 8xlarge instance in us-east-1a Create a RAID 0 volume using the four 800GB SSD
ephemeral disks provided with the instance Provision 3×1 TB EBS volumes attach them to the instance and
configure them as a second RAID 0 volume Configure synchronous, block-level replication from the ephemeralbacked volume to the EBS-backed volume.

B.
Instantiate an 12 8xlarge instance in us-east-1a create a raid 0 volume using the four 800GB SSD ephemeral
disks provide with the Instance Configure synchronous block-level replication to an Identically configured
Instance in us-east-1b.

C.
Instantiate a c3 8xlarge Instance In us-east-1 Provision an AWS Storage Gateway and configure it for 3 TB of
storage and 100 000 lOPS Attach the volume to the instance.

D.
Instantiate a c3 8xlarge instance in us-east-i provision 4x1TB EBS volumes, attach them to the instance, and
configure them as a single RAID 5 volume Ensure that EBS snapshots are performed every 15 minutes.

E.
Instantiate a c3 8xlarge Instance in us-east-1 Provision 3x1TB EBS volumes attach them to the instance, and
configure them as a single RAID 0 volume Ensure that EBS snapshots are performed every 15 minutes.



Leave a Reply 34

Your email address will not be published. Required fields are marked *


Tony Kuo

Tony Kuo

D
Creating EBS RAID 0 snapshot got to suspense all I/O until snapshot completing.

Tony Kuo

Tony Kuo

Sorry for the wrong comment.
Raid 0 cannot survive the loss of an individual disk.

Imam

Imam

Not sure why some are commenting for the sake just to comment on a question.

100,000 IOPS. Survive a loss of a disk, EC2 Instance and an entire AZ – WITHOUT data loss!

Without data loss – that alone eliminates C, D, E. No data loss = synchronous replication must be in-place.

Now we are down to A and B. A is out because it doesn’t address the failure of an AZ. In A you are replicating from instance storage to EBS. However, EBS is in the same AZ and does not address AZ outages. So the answer is B.

donkeynuts

donkeynuts

Agreed

Sail

Sail

The task is to design a storage layer. And option B says ephemeral disks.?

Phenix

Phenix

Storage does not necessary mean durable storage. Think of a cache, which also is some kind of storage. I agree with Imam that B is the only valid answer here.

Felipe

Felipe

agree.
but still wondering how to achive 100K IOPS. Max is 10K IOPS per disk
I say the question has a typo.
but then agree with B

JK

JK

I agree 100% with Imam, B fulfills all requirements. Any use of 15 EBS snapshots allows for data loss to the 15 minute interval. C has no redundancy at all.

B.

Sanjeev

Sanjeev

B.
Sustaining against loss of AZ is a key split.

Regards

hello

hello

B is not going to offer sufficient backup though?

Seems E

hello

hello

Official answer here says D, can’t be D as per AWS recommendations re: RAID5

Aneesh Mohan

Aneesh Mohan

B is satisfying all requirements except 100K IOPS,

the max IOPS it can deliver if PIOPS SSD used will be 4 x 20,000 = 80,000 which less than 100,000.

Blah

Blah

i2.8xlarge with 4x800GB SSD ephemeral can provide around 150.000 IOPS.

Blah

Blah

Sorry, i need to correct my above comment: i2.8xlarge provide about 300.000 IOPS.

Jones

Jones

who did you calculate the 300,000 IOPS?

Jones

Jones

I mean how

mutiger91

mutiger91

Probably used this:

http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/storage-optimized-instances.html

“If you use a Linux AMI with kernel version 4.4 or later and utilize all the SSD-based instance store volumes available to your instance, you get the IOPS (4,096 byte block size) performance listed in the following table (at queue depth saturation). Otherwise, you’ll get lower IOPS performance.”

i2.8xlarge
365,000 (100% random read IOPS)
315,000 (First write IOPS)

mil*

mil*

Hi Aneesh,

Can you please tell us how to count IOPS ?

Ashley

Ashley

Answer is B

Manish

Manish

It can’t be B, C or E, since all of those options doesn’t provides the data loss protection.

B is using ephemeral disks for both instance and if both instance are down then data is lost.
C There is no replication or snapshot.
E It’s set as RAID-0 so no redundancy of data and snapshot is missing 15 minute interval of data from last snapshot.
D RAID-5 provides the redundancy of data but not the IOPS or performance.

The only option left is A, whih Provides IOPS using RAID-0 and the data is synchronized at block-level replication from the ephemeralbacked volume to the EBS-backed volume. Since it’s using two RAID-0 3TB volumes one as ephemeral and one as Block Storage.

Joe

Joe

C3 are compute optimized
i2 are High I/O instances

So A or B

vladam

vladam

B is the right option.

It doesn’t protect against a loss of two EC2 instances or two AZs, but the question asks about protection of ONE disk, EC2 instance or AZ loss.

dat

dat

Agree with B is the answer.

As per I read, the question asked : “…survive the loss of an individual disk, EC2 instance, or Availability Zone without any data loss…”

A: does not provide loss prevention for EC2, although it provide enough storage capacity (4SSD*800GB > 3TB).

C: does not provide any means for loss prevention of data or ec2 or AZs
D: AWS advice not to use RAID5. And like ‘C’, D does not provide means for sustaining loss
E: using only 1 instance, no means for loss prevention of EC2.

green

green

E is correct

donkeynuts

donkeynuts

Your mom is correct

Shivani

Shivani

I donot understand how either A or B can be the answer ? .
Both A and B say – Instantiate 12 8xlarge instance – Which instance type is 12 ???
There is no such thing like 12.
I guess answer should be something between c or d or e

Harry

Harry

Answer is D
The question talks of “he storage layer must be able to survive the loss of an individual disk” and rest all options talk of Raid 0 which has no striping hence cannot withstand the data loss if one disk fails.

Anthony

Anthony

Answer is B. …if a disk is lost, you can get the data from the replicated copy in the other AZ. option D only has one AZ, so if the AZ is lost, that’s it all gone.

The question says it must survive loss of disk, ec2 instance, AZ etc.