3 months with Google Compute Cloud: Impressions

By Martin Rusev read

I have been a devoted Amazon user for a couple of years now(with a couple of DigitalOcean/Linode servers here and there) and once I heard that Google has released their own platform, competing with Amazon, I decided to use the $300 trial credit and give it a try. At this point there are more than enough awesome and reliable cloud server providers, so why would you choose Google Compute?


First up is the pricing. Before deploying my servers I directly compared the pricing between Amazon, DigitalOcean, Linode and Google Compute Cloud and Google stands out in 1 area High memory servers with Custom sized SSD.

Similar to Amazon, the little things add up and your monthly bill is not what you see in the Pricing Calculator. My experience - I deployed a 13GB RAM instance with 30GB SSD and 1 DNS zone, according to the Pricing calculator my monthly bill should have been around $80, but when I received the bill it was a little bit more than $120.

In terms of experience - the billing module is one big spreadsheet, not flexible and definitely inferior compared to the detailed Amazon Billing Dashboard where every service is grouped and you can see exactly what you are paying for.


I hope that the current interface is a work in progress and hopefully it will receive some attention in the future. As it stands right now - it is both time consuming and confusing.

The Bad

The Good

Ease of Use

With Amazon you log in to your servers with a standard pem key:

ssh -i my-key-pair.pem ec2-user@ec2-198-51-100-1.compute-1.amazonaws.com

The only inconvenience with Amazon is remembering the IP address, but if you are using zsh for example, you can automate it with aliases: alias ssh:amon='ssh -i ~*.pem ip_address'

With Google Compute Cloud, you have to install a special tool - gcloud, which needs some time getting used to. Integrating with your devops tools like Fabric, Ansible, etc could be a little bit cumbersome. The biggest benefit I saw while using the gcloud tool is better security by default - it automatically creates new unique private/public key pairs for every instance. This is something you have to do manually with Amazon, Linode or Digital Ocean.

Performance and Reliability

Once I got everything up and running - it was relatively smooth experience for the duration of my tests. I didn't encounter any slowdowns or degradation in performance which is something we are all used with VPS servers. I've encountered only one interruption for an hour or so for the whole period. With that said - it was not all great, the Compute Cloud Status page needs a lot of work. As it stands right now - it is not detailed as the Amazon Status Dashboard and the messages are little bit cryptic, generic and way to engineery for my taste.

For 40 minutes spanning Wednesday 18th and Thursday 19th February 2015, the majority of Google Compute Engine instances experienced traffic loss for outbound network connectivity, with lower levels of loss beginning at 22:40 PST on February 18th and ending at 01:20 PST on February 19th.

Takes this with a grain of salt - In the past I had to deal with the Google Support Team(for Google apps for business) and I had reported issues both to Digital Ocean and Amazon and both support teams are responsive and fast, Google on the other hand have almost non existent support and when they finally decide to aknowledge the issue, they are not very helpful. Again - my own experience, most people never have to deal with support.


My 3 months with Google Compute are up and I personally am not convinced. There are always pros and cons for every cloud provider and when it comes to Compute Cloud, these are my own: