Google Cloud Region and Zone Selection for Low Latency

CDL By islever Editorial Last updated

Computing in the cloud Operational excellence and reliability at scale

Short answer

B To achieve redundancy and less than 10 milliseconds latency, locate virtual machines in different zones within a single region.

Question

Your organization runs a distributed application in the Compute Engine virtual machines. Your organization needs redundancy, but it also needs extremely fast
communication (less than 10 milliseconds) between the parts of the application in different virtual machines.
Where should your organization locate this virtual machines?
  1. In a single zone within a single region
  2. In different zones within a single region Correct Answer
  3. In multiple regions, using one zone per region
  4. In multiple regions, using multiple zones per region

Key Takeaways

The main trap is confusing regional redundancy with latency constraints; spreading across multiple regions (Options C/D) breaks the <10ms latency rule despite offering higher availability.

This question tests Google Cloud infrastructure selection for distributed applications requiring high availability and sub-10ms latency. The correct answer is B, leveraging multi-zone deployment within a single region.

Common mistake: Selecting D because it offers maximum redundancy. This is incorrect because cross-region network hops inherently exceed 10ms latency, violating the specific performance constraint.

Expert Analysis

The core constraint in this scenario is the requirement for less than 10 milliseconds of communication latency between application components running on separate virtual machines. In Google Cloud architecture, network latency is heavily dependent on geographic distance. While zones within a single region are connected via low-latency internal networks, the latency between different regions typically ranges from tens to hundreds of milliseconds, far exceeding the 10ms threshold.

Option B (In different zones within a single region) is the correct choice because it satisfies both requirements: redundancy and ultra-low latency. Zones are designed to be independent failure domains (providing redundancy against zone-level failures) while remaining geographically close enough to maintain sub-10ms inter-zone latency. Comments supporting Option D often confuse end-user latency with component-to-component latency; however, even for internal components, cross-region links introduce significant jitter and delay that violate the strict <10ms SLA. Option A fails the redundancy requirement by placing all VMs in a single point of failure. Options C and D fail the latency requirement due to the physical distance between regions.

Official Reference

These links open the official vendor documentation in a new tab.

Exam Strategy

When analyzing infrastructure placement questions, always prioritize the most restrictive technical constraint first. If 'latency' is specified in milliseconds, eliminate any option involving cross-region or cross-continent deployments immediately, as these physically cannot meet microsecond/millisecond targets. Focus on how services like Compute Engine zones provide fault isolation without sacrificing network proximity.

Frequently Asked Questions

Can two zones in the same region have >10ms latency?

No. Zones within a single Google Cloud region are connected via a high-performance, low-latency backbone network designed to keep inter-zone latency consistently below 1 millisecond.

Why not use a single zone for lowest latency?

While a single zone offers the absolute lowest latency, it does not provide redundancy. If that specific zone experiences an outage, the entire application becomes unavailable, failing the redundancy requirement.

More CDL FAQ →

Community Votes

B
79%
D
16%
A
3%
C
2%

79% of anonymous learners picked answer B. Votes are pick records left by other test-takers — they are not the verified answer.

Discussion 91 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 30 of 91

  1. SemiDeus 👍 50
    I chose B since it guarantees a latency of ~10 milliseconds is maintained, well also giving the redundancy of multi-zone. If it was D, we could not guarantee 10 milliseconds between the different regions.
  2. prekair0 👍 12 Answered B
    Multi zone is also redundant within the region and it provides the lowest latency.
  3. Kumarmishra 👍 5
    Answer is D. Here we are not talking about latency observed by end user, it is about between components deployed on separate VMs. So lets say we choose option D, latency among components is going to be <=10 msec irrespective of in which region they are deployed. If a particular region is down then obviously end users of that region will experience higher latency as they are being served by another region.
  4. Spider888 👍 4
    The question is quite open for discussion. I would say D, as it is simply B yet multiplied by Regions. In a scale of the Region multiple Zones guaranty user access within expected 10 milliseconds, if we copy the same setup across regions all users in all regions will get the same access latency but multi zones x multi regions give you high redundancy (even if there is an issue with the region--> you just use the next closest region). What is missing in the question is some information on the coverage of the solution let's say "...Global..." or "...Regional...".
  5. xDSKick 👍 4 Answered B
    Options A, C, and D involve spreading the virtual machines across different configurations, such as single zone, multiple regions with one zone per region, or multiple zones per region. While these options may offer redundancy to some extent, they might not guarantee the desired extremely fast communication due to potential increased latency between different regions or zones. Therefore B
Show 25 more comments
  1. MHx1 👍 4 Answered B
    Single Region + Multi zone = 2 digit latency Multi Region + Multi/Single Zone = 3 digit latency Hence my answer is B
  2. grzeev 👍 4 Answered D
    you forgot redundancy! if a region is down your redundancy is dead!
  3. Dan_W 👍 3
    D . B does not provide redundancy when the region goes down. D accomplishes both objectives. Per Google docs - Zone and region failures are rare failures that can cause all of your VMs in a given zone or region to be inaccessible or fail. To mitigate these failures, create diversity across regions and zones and implement load balancing.
  4. anirban7172 👍 3 Answered C
    Answer C. for the redundancy we need to go for regions, and for low latency we have be in the same Zone. so Multiple region and one Zone per region is best choice.
  5. Matheen 👍 3
    Inter-regional Latency: triple digit milliseconds(500ms) , Inter-zonal Latency:double digit milliseconds (10ms) so therefore answer is B
  6. fpreli 👍 3
    D also for me and +1 for the poorly worded. It asks for redundancy (multi region) and low latency (multi zone), so I would with D
  7. seths 👍 3 Answered B
    latency within 10 ms seems impossible for different regions , source --> https://geekflare.com/google-cloud-latency/
  8. shamg 👍 3
    This is poorly worded Q , if we have to consider redundancy only D is best option.
  9. grzeev 👍 3
    D. we need redundancy and if region is down redundancy dead with it!
  10. ravisinha 👍 3
    Moderator can you change the answer to B ?
  11. ilberg 👍 3 Answered B
    key word 10ms...low latency. Answer is B
  12. farkallu 👍 2 Answered B
    Explanation: Low-latency communication (<10ms) requires geographic proximity between virtual machines. Using different zones within a single region ensures high availability and redundancy while keeping the network latency minimal. Google Cloud's internal network allows low-latency communication between zones in the same region, making it the optimal choice. Why the other options are incorrect? ❌ A. In a single zone within a single region – No redundancy; failure in that zone could take down the application. ❌ C. In multiple regions, using one zone per region – Inter-region communication typically has higher latency (>10ms), making it unsuitable. ❌ D. In multiple regions, using multiple zones per region – Adds redundancy but increases latency and complexity, making it inefficient for fast communication.
  13. Johno_7 👍 2
    This questions is really challenging. Sub ~10 millisecond latency would require being in the same zone. However, to achieve redundancy in case of a failure in a zone, or an entire region, we need to deploy in multiple regions. Sub 10 millisecond latency could be achieved if the entire system is replicated across multiple regions. If one goes down, users would be able to still access the hosted service but without meeting the latency requirement if redirected to the app hosted in another region. B seems safest but D also seems to make more sense in terms of redundancy. Please comment if you can add more clarity. Thanks! https://cloud.netapp.com/blog/gcp-cvo-blg-understanding-google-cloud-high-availability
  14. vikramchaganti 👍 2
    I chose D since you have the option B in D as it says there are multiple zones in a region which gives redundancy and the latency is very low within a region but incase of an entire region goes down, you need another region to serve where you will meet the redundancy factor. So, incase of B, you will have the less latency but how do you achieve redundancy incase of region failure?
  15. grzeev 👍 2
    Regions are collections of zones. Zones have high-bandwidth, low-latency network connections to other zones in the same region. In order to deploy fault-tolerant applications that have high availability, Google recommends deploying applications across multiple zones and multiple regions
  16. Monicaarg 👍 2 Answered A
    if less than 10ms is required it must be within a zone. In other words, the answer is within a single zone within a single region. To have redundancy I can work with MIG. the latency between zones within the same region is 10 ms and the question requires less than 10 ms.
  17. JK1111 👍 2
    Can’t understand how some of the discussions are made here.. the answer is B Check Mr Peter Browns video about regions/zones where he specifically talks about this question. If you want to achieve 10ms latency you can’t have multi regions. End of discussion
  18. jasbans 👍 2
    D is correct as per Google: for apps such as real-time games, a few milliseconds of latency can have a greater affect on user experience. Deploy these types of apps in multiple regions close to the users.
  19. aasgtr 👍 2
    Latency between zones is 10ms while latency between regions is 500ms. The question specifically states 10ms so answer must be B. Multi zone offers redundancy.
  20. adyanrado 👍 2 Answered D
    D is correct because guarantee slow latency between regions and zones in case of any zone failed, yet have slow latency between regions
  21. PHD_CHENG 👍 2 Answered D
    D is correct. You are running distributed application. Multi region multi zone provides not only better latency, but also, resilience
  22. AnithaPal 👍 2
    B as the latency required is less than 10 ms, for inter regional latency may be triple digit milliseconds
  23. Bhailog 👍 2
    Its D. InterRegional Latency is 3 Digit ms. Inter-Zonal is 2 Digit. 10 millisecond.
  24. anverdor 👍 2
    The latency betweem regions is around 3digits and inter-zones 2 digits. So its B
  25. NU_1234 👍 2
    D - multiple regions offer redundancy and multiple zones offer low latency

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Public Cloud Provider Ownership in Shared Responsibility Model

S-grade

Google Cloud Data Lake vs Data Warehouse Selection

S-grade

Migrating Private Cloud Hypervisor Workloads to Google Cloud

S-grade

Strategic Data Analysis Service for Petabytes

S-grade

Practice All CDL Questions

Access 84 questions with complete answers and detailed explanations.

View Full CDL Practice Test →

← Back to CDL Study Guide