Latency ‘map’ for GCP performance dashboard

A pre-read of my earlier blog on this subject may be useful

GCP Performance dashboard is a module under ‘Network Intelligence Centre’ which focuses on GCP Network related performance statistics . Performance Dashboard gives our customers visibility into the performance (latency and packet loss ) of the entire Google Cloud network, as well as to the performance of your project’s resources. One of the cool features of the performance dashboard is the visualization of network performance on a world map . This blog gives more details of this map feature and how this can be useful .

To start off , the map feature of performance dashboard can be located under following

1. Google cloud Latency → Google Cloud to Internet endpoints RTT → Map
2. Latency → Google Cloud to Internet endpoints RTT → Map

The difference between a) and b) above is the fact that option a) gives latency numbers for the whole of Google cloud irrespective of whether the customer has workloads deployed in GCP or not . Option b) gives details very specific to customer projects and hence will allow customers to list the regions where they have active workloads . The below snip gives an example of Google cloud latency

2 symbols are to be noted in this map →

This symbol signifies a region of Google cloud

This symbol signifies a city on global map

Equally important are the colors of these symbols . Let me share some details on the same :

The region symbol can have one of the above 2 colors . The color gray signifies that the user has not selected the region on the map . If user selects one of region (so that details specific to that region needs to be shown ) , the region is shown with a violet color

The color of a circle showing a given city on a global map can be one shade of Blue as per above guidance . The lighter the shade of blue , less is the latency of that city from a given GCP region.

Now , let me take help of a few use cases and depict how the performance dashboard map may be useful:

Use case # 1: Latency from a city to a GCP region

Let’s take an example where a customer is thinking of hosting his application in GCP region asia-southeast1 (Singapore). Majority of users of this application are in India . Customer want to know a range of latency that users may see when they try to access the application

Since the user is ‘planning’ to host the application but hasn’t still hosted anything in his account , we need to go and look for required data under “Google cloud Latency” . Once we are on this page , select the region for which we are seeking details . In our example , we selected the “asia-southeast1” region . The selection is shown in below image –

• At a very first sight , from the shades of blue , a user can make a high level judgment about latencies from India to the GCP Singapore region.
• Lets say , there is a requirement which requires us to find latencies from 2 specific Indian cities (Bangalore and Kolkata ) to GCP asia-southeast1 region . For this , a user has to hover on / click the cities to find the latencies . Example -

• Clicking on “View timeline” will help a user to find the required latency over a period of last (upto) 6 weeks . Example

Use case # 2 : Find the list of nearby regions

Let’s take another example . A customer has a majority of users in or around Helsinki, Finland and wish to find out hosting location(s) which are nearby and have least latencies from customer location .

Since in this case also , the customer has no application in his GCP project. So the user must navigate to “Google cloud Latency” tab.

• Ensure that no GCP region is selected . i.e. all the regions on map are shown in Grey color diamond .
• In the search section on top , search for location [ Example : Helsinki , Finland ]

• As a result , user sees GCP regions which are nearby and also mentions the latency of Helsinki city from the GCP regions
• If a user want to find , trend of latency numbers in last 6 weeks , he can do so by going to “View Timeline”

Use case # 3 : Latency numbers for Premium vs Standard tier

Assuming that the customer has decided to host an application in asia-south1 (Mumbai) , he wish to understand what will be latency for his clients in US-west coast over Standard tier and Premium tier network .

Closing Notes:

With the help of ‘Performance dashboard’ users of Google cloud have a great toolkit in their hands to understand the network performance statistics . The ‘map’ feature is a cherry on the cake . Instead of searching source as city X , destination as GCP region Y for Google cloud to Internet endpoint RTT , the map feature gives a global view of how latencies look from a global lens . The map presents more intuitive view of latency between various global cities and a GCP region and hence can be a good tool to judge a range of expected latency between a city and a GCP region

Disclaimer: This is to inform readers that the views, thoughts, and opinions expressed in the text belong solely to the author, and not necessarily to the author’s employer, organization, committee or other group or individual.