Getting started with Cloud Computing using R Programming

Introduction

Almost any domain / business today is being transformed through SMAC. SMAC is a collective term referring to changes happening in Social, Mobile, Analytics and Cloud. The impact of this change has been across the spectrum – Organizations, people and Products. In today’s article, we will enable you to take your analytics capabilities to next level by using Cloud computing.

We have explained the concept of cloud computing using R programming and RStudio using a step-wise methodology. Furthermore, you will also learn about the benefits of using R over cloud as compared to the traditional desktop or Local client / Server architecture.

Cloud – an enabling platform for data science:

Cloud computing has witnessed an unparalleled growth and penetration in last few years. It has enabled organizations to scale quickly and easily. Using cloud services, companies today collect, store and analyze huge amount of data, which was almost non-thinkable before. However, with services from the likes of Amazon, Google and Microsoft, cloud services are now accessible to any analyst.

Gone are the days, when you purchase a server for a particular capacity and then need to purchase a new one, when you grow out of the previous capacity. For example, most of the analysis I normally do is on a few GBs of data – sufficient to run on my laptop directly. However, recently Microsoft released ~400 GB of data about Malware and viruses on Kaggle. If, I would have thought of solving this problem on my laptop, I would have run out of my internet plan in just downloading the dataset. Analyzing it is a separate challenge in itself.

Even if I would have downloaded the dataset, the only way to do meaning computation through non-cloud way was by buying new machine – not a very practical solution. This is where cloud computing comes in picture!

Must Read: Step wise guide to learn R Programming

Why do you need the ‘cloud’?

As discussed in the case study above, cloud is cheaper for handling big data than storage on local desktops, laptops or servers. Wait. Big Data?  Yes! Big Data is an umbrella term that basically denotes data whose Volume and Variety and Velocity is larger than conventional data sources and which requires distributed computing like Hadoop and non-RDBMS storage like NoSQL databases.

Must Read: A beginner’s guide to use big data using MongoDB

What is Cloud Computing?

According to the NIST definition of cloud computing,

Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model is composed of five essential characteristics, three service models, and four deployment models.

Cloud Computing consist of 3 components:

• Infrastructure as a Service(IaaS)
• Platform as a Service(PaaS)
• Software as a Service(SaaS)

IaaS– To deploy their applications, cloud users install operating-system images and their application software on the cloud infrastructure. In this model, the cloud user patches and maintains the operating systems and the application software.

PaaS– Cloud providers deliver a computing platform, typically including operating system, programming language execution environment, database, and web server. Application developers can develop and run their software solutions on a cloud platform without the cost and complexity of buying and managing the underlying hardware and software layers

SaaS – In software as a service (SaaS), users are provided access to application software and databases. Cloud providers manage the infrastructure and platforms that run the applications. SaaS is sometimes referred to as “on-demand software”.

What are the cost benefit trade-offs of using cloud computing with R versus other applications?

Python is free just like R, but the main reason R scores is that the statistical library of R packages is far more extensive. SAS remains the leading language for corporate analytics on the desktop but it remains expensive for small enterprises and has a significant disadvantage in capital expenditure commitment because of annual license structure instead of one time licence fee.

Must Read: A Quick Guide on SAS vs R vs Python

What are the advantages of using R on the cloud versus on the desktop?

1. Since we know R is constrained to handle data only as big as the RAM size, the cloud offers us a quick solution to handle Big Data Science using R. This can be done by simply ramping up the RAM on the virtual machine instance.  You can see the various kinds of RAM options available on the cloud which are simply not affordable  on the local machine.
2. For big Datasets , it is better to use it on the cloud than to download the dataset, process it and then score it. For example, if you have a competition that uses 30 GB of data, it is best you use it on the cloud. The cloud is thus a great way of learning about Big Data without getting hassled by internet speed.
3. Cloud has much better bandwidth speed. Therefore, installing software and transferring data is much faster on the cloud.
4. You can use additional services like AzureML with R on the cloud rather than build your own machine learning service from scratch. You can this tutorial for more information.
5. Cloud is much more scalable for changes in volume or velocity of data.

Take the Test: Should I become a Data Scientist?

How to use R Programming on the cloud?

You can create a instance (a virtual machine that you access remotely) on Amazon Cloud, or on Microsoft Azure or on Google Cloud. You can then simply install R the same way as you use it on your local desktop. You connect to your remote machine through SSH or Remote Desktop.

Here is a step by step process for creating a cloud instance on Amazon Web Services.

Note: Amazon has a free tier that enables you to try out the Amazon cloud for free for 1 year. However this is only for micro instances which have very small RAM and very small disk space. For higher RAM and higher storage you need to pay more. To look at the various instances and their per hour pricing you can see visit here. Basically fees is charged in compute units but this website makes it easy to figure out costs.

First you need to create your Amazon Id. Once you are done, follow the steps below to create a cloud instance on amazon web services:

1. 1. Login to Amazon Web Services (AWS) Console
   2. Click on Run Instance
   3. Choose operating System for your virtual machine that you will access remotely. Here I have choosen Amazon Linux.
   4. Choose Instance Type (size of RAM  and memory needed). See here to compare prices.
      
   5. Create a  security key. This is needed for a secure cracker-proof login to the remote machine. Note you can use remote desktop for Windows operating systems but you will need to use SSH for Linux Instances.
   6. Click on Launch  Instance
   7. Connect to the instance using your security key following the instructions given.
   8. Now work on your remote machine just as you would work on a local machine.
   9. Here I am trying to install R                                                                  
   10. Once your work is done- remember to please close the instance lest you incur high monthly bill.

You can choose on demand instances, or even have reserved instances (booking a virtual machine for a fixed period of time and thus at a considerable discount).

Take the Test: Should I become a Data Scientist?

How to use R on the cloud using RStudio?

RStudio Server edition runs on only Linux. Therefore, we choose  Linux instance on the cloud and then configure R Studio Server. We can then connect to the remote RStudio Server through the browser and use it just the same way.

Here is a step by step way to run RStudio on the cloud.

• Note- We installed R already using sudo yum install R
• Download the RStudio server to your virtual machine and then install RStudio Server

$ wget http://download2.rstudio.org/rstudio-server-rhel5-0.99.442-i686.rpm
$ sudo yum install --nogpgcheck rstudio-server-rhel5-0.99.442-i686.rpm

• You verify the installation

$ sudo rstudio-server verify-installation

• You open the 8787 port using the security group in AWS Console(left margin Security Groups) by creating a custom TCP rule (click Edit on the tab below)

• You create a new user with a new password using the SSH terminal for your virtual machine cloud instance

1. sudo useradd newuser1
2. sudo passwd newuser1

• You find the public IP Address of the cloud instance from the Instances tab in the left margin

• You open your browser to IPAddress: 8787 and then login using the user id and password created above

• Now you are ready to use R using the cloud through a browser

Using R through the Bioconductor cloud?

Bioconductor cloud is an amazing way of kickstarting R on the cloud. You can see the instructions here.

What are the various other cloud computing options?

You can use cloud options from Google and Windows Azure as well. However most of the space is dominated by Amazon Web Services.

Any examples of Using R with Platforms and Other software as a service?

Yes we can use Azure Machine Learning with R on the cloud and also use Google Big Query with R.

Any examples of Big Data using R on the cloud?

Yes there are many examples. Resource 1 and Resource 2.

End Notes

By now, you would have got an overview of how to implement cloud computing using R and R Studio. I really enjoyed writing and curating the useful resources in this article. This article also covers questions which are usually asked by people while learning cloud computing in R. Hence, I have tried to cover all of them in this article. As per my personal experience, I’ve found demonstrating cloud in R is relatively easier as compared to other softwares.

I hope this article helped you to become familiar with cloud computing. We would love to hear about it from you. Did you find it useful? Feel free to post your thoughts through comments below.

If you like what you just read & want to continue your analytics earning, subscribe to our emails, follow us on twitter or like our facebook page.