Sometimes you have a data analysis problem that is just too big for your desktop or laptop. The limiting factor here is generally RAM. Thankfully, services like Google Compute Engine allow you to lease servers with up to 208GB of RAM, large enough for a wide variety of intensive tasks. An ancillary benefit of using a service like Compute Engine is that it allows you to easily load your data from a Cloud Storage Bucket, meaning you don’t need to keep a copy of the large dataset locally at all times.
R Studio has a remote mode allowing you to install it on a server with access through a remote interface. This tutorial details how to start a Compute Engine instance, install R Studio on it and access R Studio from the remote interface.
The rest of this tutorial assumes that you have a Google Cloud Platform account with billing enabled and have installed the Google Cloud SDK.
Deploying a Compute Engine Instance
The first step is to deploy your Compute Engine instance. The
command allows you to create instances. The only required parameter to create an
instance is the instance name. We will call our instance
r-studio but you can
choose any name you like. R Studio Server is typically built on Ubuntu so it is
safest to use the Ubuntu distribution for your server.
gcloud compute instances create r-studio
You will be prompted to choose a
Zone. Just choose a zone close to
you. You can also specify the zone when creating the instance using the
parameter. For example.
gcloud compute instances create r-studio --zone us-central1-a
You will also have to open the Compute Engine firewall to allow port 8787 for R Studio.
gcutil addfirewall allow-r-studio --allowed=tcp:8787
Installing R Studio
Once we have our Compute Engine instance set up, we log in to the machine using ssh.
gcloud compute ssh r-studio --zone us-central1-a
Now that we are logged in to the Compute Engine instance, it’s time to install R by first updating the Debian apt-get repository and then installing R.
sudo apt-get update sudo apt-get install r-base r-base-dev
R Studio currently requries OpenSSL version 0.9.8. We need to install this separately and then install install R Studio
wget http://ftp.us.debian.org/debian/pool/main/o/openssl/libssl0.9.8_0.9.8o-4squeeze14_amd64.deb sudo dpkg -i libssl0.9.8_0.9.8o-4squeeze14_amd64.deb sudo apt-get install gdebi-core wget http://download2.rstudio.org/rstudio-server-0.98.1103-amd64.deb sudo gdebi rstudio-server-0.98.1103-amd64.deb
You should be up and running with R Studio on your compute engine instance. To verify, navigate to the IP address of your Compute Engine instance on port 8787 (the default R Studio port).
R Studio only permits access to users of the system, we can add a user with standard Linux tools like adduser. For example, to create a new user named rstudio and specify the password you could execute the following commands.
sudo adduser rstudio
You will be prompted to enter a password for the user and confirm the users name and phone number.
Afterwards, logging in with the user you created will present a web UI of the familiar R Studio. You can now perform analysis on those larger data sets using the R Studio that just weren’t possible on a laptop.