exam questions

Exam Professional Cloud Architect All Questions

View all questions & answers for the Professional Cloud Architect exam

Exam Professional Cloud Architect topic 11 question 3 discussion

Actual exam question from Google's Professional Cloud Architect
Question #: 3
Topic #: 11
[All Professional Cloud Architect Questions]

Dress4Win has asked you to recommend machine types they should deploy their application servers to.
How should you proceed?

  • A. Perform a mapping of the on-premises physical hardware cores and RAM to the nearest machine types in the cloud.
  • B. Recommend that Dress4Win deploy application servers to machine types that offer the highest RAM to CPU ratio available.
  • C. Recommend that Dress4Win deploy into production with the smallest instances available, monitor them over time, and scale the machine type up until the desired performance is reached.
  • D. Identify the number of virtual cores and RAM associated with the application server virtual machines align them to a custom machine type in the cloud, monitor performance, and scale the machine types up until the desired performance is reached.
Show Suggested Answer Hide Answer
Suggested Answer: D 🗳️

Comments

Chosen Answer:
This is a voting comment (?). It is better to Upvote an existing comment if you don't have anything to add.
Switch to a voting comment New
MyPractice
Highly Voted 4 years, 7 months ago
A - not correct. as its talking about Physical server size B - not correct. as we its talking about Max spec C - not correct. as its talking about the Smallest spec D - is CORRECT. as its recommending to map with on premises app VM Size
upvoted 41 times
Smart
4 years, 5 months ago
Agree. Starting with a minimal size at which your application can run efficiently/optimally is the best practice. This would be best estimate from past usage so monitor closely and apply vertical and horizontal scaling.
upvoted 3 times
...
Ziegler
4 years, 2 months ago
What if the on-premises workloads are oversized? C is the correct answer instead
upvoted 4 times
JMSTP
2 years, 9 months ago
Agreed, but confused as question states they're moving Test/Dev first and C says "into Production". If they're looking to save money, start small and grow until performance needs are met. Test/Dev is typically tolerant to these incremental changes.
upvoted 1 times
...
Rafaa
4 years, 2 months ago
They have scaling problem, hence decided to move to GCP.
upvoted 3 times
rottzy
2 years, 10 months ago
Also, cost-cutting was a primary concern = while moving to cloud
upvoted 2 times
...
...
...
...
jcmoranp
Highly Voted 4 years, 9 months ago
It's D. You can monitor machines to scale until necessary
upvoted 25 times
techalik
3 years, 8 months ago
D i think: Start with the smallest instances and scale up to a larger machine type until the performance is of the desired standard. is the right answer. In continuation of the above explanation, although you could use a predefined machine type like e2-highmem-4/n2-highmem-4/n2d-highmem-4 etc. if you need 4 VCPUs and 32GB memory, there's no guarantee that it performs similar to the existing VM in the data centre. The networking fabric is different, the disk I/O is different, and the CPUs are different too. We don't know the exact specifications of the data centre CPUs to draw a parallel to the processors offered by GCP. As you can see, the performance can vary a lot depending on the frequency. (remember shelling out additional 500$ for upgrading CPU from 2.6GHz to 2.8GHz when buying your a laptop?). You may realize that you need more vCPUs after migrating to Google Cloud or maybe less, but until you migrate and test it out, there is no way to say which is the best machine type. So the recommendation should be to start small, increase the instance size as needed until the performance is of an acceptable standard, and that is your machine type.
upvoted 3 times
techalik
3 years, 8 months ago
I meant C :) explanation is for C
upvoted 3 times
...
...
nitinz
3 years, 5 months ago
C is the best answer. Thats the beauty of cloud. You can change machine type by just shutting it down. Also Google discourages using custom VM sizing.
upvoted 2 times
...
Pankonics
3 years, 7 months ago
C is correct. In option D, it says No.of virtual cores and RAM associated... Which is not mentioned in Case study as well. This option just trying to confuse that set. So will go with C.
upvoted 1 times
...
chiar
4 years, 8 months ago
Remember that the VM's CPU RAM and Disk are designed in GCP for optimized performance, I think C is the best
upvoted 3 times
MyPractice
4 years, 7 months ago
the smallest instance avail is as low as 1CPU with 128MB RAM - Is that sufficient for prod run (assume currently its running with 24CPU with 128GB of ram in on premise)
upvoted 5 times
Rathish
4 years, 2 months ago
I thought, only dev and testing environment migrated first and in that case, I prefer to start with smallest instance available. - Dynamic scalability is the reason, we are moving to cloud.
upvoted 2 times
kk4gcp
3 years, 5 months ago
but the option c says - C. Recommend that Dress4Win deploy into production - they are talking about production - so D would be correct
upvoted 2 times
...
...
...
...
...
user263263
Most Recent 1 week ago
Selected Answer: D
The best solution would be "use a small machine type suitable for the app server, use MIGs and autoscaling" but that's not in the answers. B - no, the legacy app servers are 4 CPU, 32 GB (1 CPU : 8 GB), high memory machine types are 1 CPU : 30 GB (e.g. m3-ultramem-32) C - no, the smallest instance available is too small for an app server A or D - It depends on "is the on perm env virtualized?". In "Storage appliances" the case study mentions "iSCSI for VM hosts". If that applies to the app servers, the correct answer is D.
upvoted 1 times
...
bigzero
2 months, 3 weeks ago
Option A D can do after migration finished in production environment before migration, you can only do the option A
upvoted 1 times
...
hynglly
7 months ago
Selected Answer: D
it is D,
upvoted 1 times
...
MahAli
7 months, 3 weeks ago
Selected Answer: A
Application servers as mentioned in the study, doesn't mean VMs, nothing is mentioning VMs on prem, besides custom types on GCP cost more than predefined types, I'll map their physical servers to the nearest machine type.
upvoted 3 times
...
luke19962023
1 year, 3 months ago
Selected Answer: D
since they have VM's in current environment, I'll choose D. This lets you like for like replace the on prem footprint and scale from there. I originally thought A, however, using custom machine image will prevent you from paying for running a larger machine image than necessary.
upvoted 1 times
...
taer
1 year, 4 months ago
Selected Answer: D
By identifying the number of virtual cores and RAM associated with the application server virtual machines and aligning them to custom machine types in the cloud, you can create an environment tailored to Dress4Win's specific needs. Monitoring performance and adjusting the machine types accordingly ensures that the infrastructure is optimized for both performance and cost.
upvoted 1 times
...
BeCalm
1 year, 5 months ago
Selected Answer: A
D is an iterative approach to finding the right specs which is definitely not the way it works IRL.
upvoted 1 times
...
stevehlw
1 year, 8 months ago
I'll be truly surprised if the correct answer is C. Deploying the smallest instances into production will likely trigger a ton of error messages in real life. If you're oversizing the VMs, just simply scale it down.
upvoted 2 times
...
jabrrJ68w02ond1
1 year, 9 months ago
IMPORTANT: Dress4Win is not anymore part of the officially listed case studies: https://cloud.google.com/certification/guides/professional-cloud-architect
upvoted 5 times
...
AhmedH7793
1 year, 10 months ago
Selected Answer: D
It is D
upvoted 1 times
...
ShadowLord
1 year, 10 months ago
Selected Answer: A
https://cloud.google.com/migrate/compute-engine/docs/4.9/concepts/planning-a-migration/cloud-instance-rightsizing 1. Performance-based recommendations: Recommends Compute Engine instances based on the CPU and RAM currently allocated to the on-premises VM. This recommendation is the default.
upvoted 5 times
AlizCert
1 year, 2 months ago
I agree with the thought process, but A is about the _physical_ HW that hosts the on-prem VMs, not about the VM requirements, so it should be D.
upvoted 1 times
...
...
[Removed]
2 years, 4 months ago
A should be better. https://cloud.google.com/architecture/resource-mappings-from-on-premises-hardware-to-gcp
upvoted 3 times
Matalf
2 years ago
Resource Mapping: "In a scenario where your applications are running on bare-metal " Rightsizing: "migrating a virtual machine to Compute Engine"
upvoted 1 times
...
...
OrangeTiger
2 years, 7 months ago
Selected Answer: D
I vote D. https://cloud.google.com/migrate/compute-engine/docs/4.9/concepts/planning-a-migration/cloud-instance-rightsizing
upvoted 2 times
...
MF2C
2 years, 7 months ago
Selected Answer: D
vote D
upvoted 1 times
...
atlasga
2 years, 7 months ago
Wow, these answers are so dumb I'm compelled to comment so new people don't mistakenly think this is how it works in the real world. You're not supposed to map sizes one-to-one OR start with the smallest instance sizes available. The industry best practice is to right-size based on actual utilization. I guess I'll pick C if this question comes up, because at least it will be cheaper than starting with big instances (but it could cause production operational issues if the instances are undersized!).
upvoted 2 times
theBestStudent
8 months ago
but then, you don't want to cause production issues for wrong sizing. If so, it ca not be C in my opinion. The main idea is always avoid prod issues
upvoted 1 times
...
...
Community vote distribution
A (35%)
C (25%)
B (20%)
Other
Most Voted
A voting comment increases the vote count for the chosen answer by one.

Upvoting a comment with a selected answer will also increase the vote count towards that answer by one. So if you see a comment that you already agree with, you can upvote it instead of posting a new comment.

SaveCancel
Loading ...
exam
Someone Bought Contributor Access for:
SY0-701
London, 1 minute ago