Evaluation of Amazon's Cloud Computing

Table of Contents

Summary

Introduction

Category 1: Compute

a.  Amazon Elastic Compute Cloud (Amazon EC2)

b.  Amazon Elastic Load Balancer (ELB)

c.  Amazon Lambda

Category 2: Database

a.   Amazon Relational Database Service (RDS)

b.  Amazon DynamoDB

c.  Amazon Elastic Cache

Category 3: Internet of Things

a.  AWS Internet of Things (IoT)

Category 4: Analytics

a.  Amazon Machine Learning(AML)

b.  Amazon Data Pipeline

c.   Amazon Elastic Map Reduce

Conclusion

References

Summary
 

In today’s technology-driven world ‘cloud computing’ is at boom. It is attracting companies globally with its unique characteristics like scalability, cost effectiveness, no upfront commitment. The AWS Cloud provides a collection of infrastructure services, namely computing, database, IoT, and analytics that are delivered on-demand, available in few clicks, with pay-as-you-go pricing. No upfront capital expense is needed in the addition of new services. AWS provides services suitable for any organization whether small or large. Launched in 2006, the AWS platform adds new services and regions (data centres) continuously since then. This report gives an overview of AWS Cloud and services which constitutes the platform.

Introduction

Cloud computing refers to services and solutions that are delivered and consumed in real time over the internet. Cloud technologies are widely used by companies like Netflix, Instagram, Dropbox, Airbnb, & many more. The below three aspects are new in cloud computing: –

      On demand availability of unlimited computing resources, thus no advance planning is required for provisioning

      Allows companies to start with small scale & later on scale in or scale out hardware resources as per requirement

      Pay as you go model – You only pay for what you use, making cloud services as cost effective

There are many cloud providers like AWS, Microsoft Azue, IBM Cloud, Racksack, etc. However, AWS is the leading cloud provider worldwide recognized by Gartner group & Forrester Research. AWS has not only marked its presence over businesses but also over space. NASA’s Mars Mission was built using by AWS services to stream the images and videos of its spacecraft landing.

Amazon Web Service (AWS)[1] “is a secure cloud services platform, offering compute power, database storage, content delivery and other functionality to help businesses scale and grow.” Initially it was built to handle Amazon’s internal online retail operations but soon it became one of the first companies to present a pay-as-you-go cloud model that is scalable.
AWS is majorly infrastructure as a service (IaaS) but is comparable to other offerings like SaaS & PaaS (software/platform as a service).

 
AWS portfolio comprises of more than 100 services categorised into functionalities.
Below four functionalities are discussed in details in this report–

1. Compute
2. Database
3. Internet of Things (IoT)
4. Analytics

Category 1: Compute
 

a.      Amazon Elastic Compute Cloud (Amazon EC2)
 

Amazon EC2 helps to obtains virtual services which are known as computing instances on demand and within minutes. This helps in creating & deploying applications quickly. The instances can be scaled up or down using Amazon EC2 management console or automate the process using API. The instances can be present in same or different geographical regions, known as availability zones (AZ) which are connected by low latency network.

Companies like Airbnb[2] is using AWS platform from day one. At present Airbnb is using more than thousand EC2 instances & they are growing over time. Airbnb has saved the cost of almost one operations team by using AWS services. Ever since company migrated to AWS it was benefitted by the features of it. It can now handle the high demand rush more efficiently.

Benefits:

● Reliable – instances can be quickly replaced & provides 99.9% availability of instances as per SLA for each Amazon EC2 region.
● Design – works well with various AWS services like Amazon S3, Amazon RDS, Amazon DynamoDB, etc
● Secure – works in Amazon VPC for security & robustness
 

b.      Amazon Elastic Load Balancer (ELB)
 

Elastic Load Balancer automatically disperses the load evenly to numerous instances. Whenever client sends a request to it, the request is redirected to one of the available instances.

Researches indicates that IT companies like Wipro[3] which is basically service company faced a lot of challenges at service helpdesks initially. With high incoming traffic load of around 200,000 tickets each month it was difficult for company to handle them quickly & promptly.
Kavitha Kadambi, general manager at Wipro, put it herself, “A cloud platform would give us the flexibility to scale capacity on-demand. It would also improve speed of deployment, enhance availability, and reduce operational cost.” (Wipro Case Study – AWS)
Today, Wipro has moved to AWS solution & its services. It uses ELB and Auto Scaling to distribute the incoming traffic to EC2 instances and scales them

automatically when needed. This solution has made the life of Wipro employees much easier & comfortable.

Benefits:

• Elastic ­­-    scaled of handling sudden changes in network traffic patterns
• Secure ­-    security groups for the load balancer, which means you can allow traffic as your wish & rest everything is blocked.
• Integrated– integration of 14 different AWS services

c.      Amazon Lambda
 

AWS Lambda is a compute service which runs backend codes(functions) and response to actions such as objects upload to amazon S3 bucket, updates to dynamo DB tables, or in app activity. It allows developers to focus on developing applications without worrying about scaling, capacities, patching & administration of infrastructure to run the code and provides visibility in the performance by publishing real time logs to amazon watch. It allows applications to deliver great user experience like live streaming, rapidly displaying street address or image icons by responding to actions within applications. It is very cost effective. It allows user to pay only for the time the code is running measured in increments of 100ms.

Netflix[4] was the first users of AWS Lambda. It provides 10 billion hours of videos to millions of subscribers across world. In order to serve huge audience Netflix moved completely from its own data centres to AWS cloud infrastructure. Netflix is using AWS Lambda to create rule-based triggers to automatically encode large media files, backup validations, deployments of instances, and monitoring the AWS resources. This helps in removing inefficient processes to reduce error rates and saves time.

Benefits:

● Any existing Lambda functions can be used which is defined in workflows
● No need to design a program to implement and execute them as Amazon SWF directly calls lambda functions
● Lambda provides us the metrics and logs for tracking function performance.

Category 2: Database
 

a.     Amazon Relational Database Service (RDS)
 

Amazon RDS allows to setup, operate & scale new databases behind the scene on single click thus providing durability, accessibility & availability to the data . There is no need to buy, rack & stack hardware, and no need to install software. In case of disaster event, RDS will switch over to the standby and you can replicate the data to another AWS region.

Amazon RDS has influenced hotel industries as well. One of the world’s largest hotel, Choice Hotels[5]International, Inc. with 6,800 hotels franchised in more than 40 countries and territories. According to Brian Kirkland (Chief Technology Officer, Choice Hotels), the hotel has seen about 300% increase in the volume of booking request over last 3 years. This made them to significantly adopt use of Amazon Aurora (Amazon RDS database engine) & other AWS services to develop new booking reservation system. It enhanced the performance and worked out to be very cost effective for the hotel.

 Benefits:

● Host replacement – in case of break down, it automatically switches to standby.
● Automatic backups – backups everything automatically for recovery in case of failure
● Software patching – apply latest patches to DB automatically using DB Engine Version Management

b.     Amazon DynamoDB
 

It is a high performance, non-relational database service built for cloud. It is easy to use & it is very fast. It uses SSD (solid state drive) technique to store data which provides fast access to the data. It is very reliable as it automatically replicates data across multiple AWS availability zones to protect data.

Researches shows that online learning sites like Duolingo[6] uses Amazon DynamoDB
to store 31 billion items which delivers lessons for 80 languages. The start-up company has reached about 18 million users globally who perform billions of free Duolingo exercises. It uses DynamoDB as it is highly scalable, reliable & proves

greater performance of approx 24,000 read units/sec and 3,300 write units/sec.Duolingo also uses other AWS services like EC2, Elastic Cache, S3, RDS. Going forward, Duolingo is planning to use further complex services like AWS Lambda & Redshift.

Benefits:
 

● Managed service - No need to install SQL or configure distributed database cluster. In order to meet I/O performance requirements, it scales, handles, and re-partitions data over more machine resources.
● Durable and available – It replicates data over different availability region for disaster recovery
● Flexible -allows creation of dynamic tables including multi-valued attributes.
 

c.  Amazon Elastic Cache
 

It is a web service which helps to improve performance of web applications.
It allows users to retrieve information from fast, managed, in memory caches instead of slower disk-based databases. It can significantly improve latency and throughput for many heavy read applications.

PlaceIQ[7], is a company which provides information about type of people present in a particular location at a particular time. Basically, it is a location intelligence company.

Therefore, there is a need to process huge amount of data at backend & handle real time challenges. PlaceIQ turned towards Amazon ElastiCache as “Amazon ElastiCache provided a turn-key approach to installing, managing, and scaling a large Memcached cluster,” says Steve Milton,CTO and co-founder of PlaceIQ(AWS Case Study: PlaceIQ).The company showed revenue profit upto $1000 per month & performance improvement by 83 after switching to AWS solution.

Benefits:

● Performance -stores data in-memory and cache to provide quick responses to applications.
● Fully managed – automatically monitors the underlying infrastructure for hardware provisioning, software patching, setup, configuration, failure recovery,etc
● Scalable – It can increase or decrease DB instances as per requirement

Category 3: Internet of Things
 

a.      AWS Internet of Things (IoT)

AWS IoT is a cloud platform that lets connected devices- car, light bulbs, sensor grid and more to interact easily & securely with cloud applications and other. All the AWS IoT services are serverless that means they can take full benefit of elasticity of the AWS Cloud & are scalable. AWS IOT provides edge-based service & cloud-based services.

AWS edge-based services are AWS FreeRTOS & AWS Greengrass & has below functions : –

• connect the devices securely
• Gather data
• collect intelligent actions locally, even when connectivity is down

AWS cloud-based services are AWS IoT Core, AWS IoT Device Management, AWS IoT Device Defender & has below functions: –

• Quickly onboard large and diverse group of devices
• Maintain group health
• Keep the group secure

On top of this we can analyse the IoT data using AWS IoT Analytics technology, having below functions: –

• Easily analyse IoT data
• Integrates seamlessly with Amazon QuickSight for visualisation and Amazon SageMaker for hosted machine learning

Startup company like Zimplistic[8] has made Rotimatics, a smart device which can make fully baked rotis in less than a minute. Rotimatics has earned more than $20 million from revenue and a total of 10 million rotis have been made ever since it is launched. The success of Rotimatics goes to IoT capabilities of AWS which improved the performance of device by responding to devices which are connected to it & troubleshoot. More user data is available now about their usage patterns, gauge feedback, satisfaction level & also about favourite recipe. With AWS we pay as we consume the cloud resources, keeping our IT cost aligned to business growth.

As said by Rishi Irani (Co-founder & CEO Zimplistic), himself” The market for smart devices in the home is flourishing, and AWS IoT services are helping us be part of that development by supporting our Rotimatic flatbread-making robot” (Zimplistic Case Study-AWS)

Category 4: Analytics
 

a.  Amazon Machine Learning(AML)
 

AWS AML service helps to develop smart applications using predictive analysis which are based on user’s data. Machine learning technology identifies patterns in data and build mathematical model using these patterns. So that the predictions on new data can be made using these models. It uses “industry-standard logistic regression” algorithm to generate models. These applications can perform important function like demand forecasting, predictive customer support, fire detection and quick decision. In addition to this we can use AML for email analysis, product reviews, forms or fund transfer and recommend correct actions to different teams.

Started in 2008, Buildfax[9] worked on aggregating distributed building permit data across US.The biggest customer base of Buildfax is insurance group which spends annually billions of dollars on roof losses. Insurance company depends on data from buildfax for creating policies & premiums.

Initially, buildfax used model based on ZIP codes, Python and R languages to estimate the age & conditions of roofs. It was a complex approach & doesn’t produces desirable results. Hence, they switched to AML which was easier, faster & more accurate solution. They used data from public sources and customers data to build models. The company can now build models within few weeks as well as became capable of providing 80% accurate results.

Benefits:

● Easy to create machine learning models from data stored in amazon databases and using wizards & API’s.
● High performance – capable of providing billions of predictions for the applications.
 

b.  Amazon Data Pipeline
 

It is a web service, integrates data from multiple AWS services and provides a single location for analysing it. This way it is easier for users to work on data.

Salesforce DMP[10] is the renowned data management platform which captures, unify and activate signature data across various devices and sources like social, videos, blogs in real time. Therefore, it interacts with billions of data monthly to provide best customer experience. Due to this, the company faced the challenge faced of quickly & efficiently processing large amount of data. Salesforce DMP moved to AWS for data processing which involves real time, on-demand, batch mode, on-demand analysis of petabytes of data. It uses AWS Data Pipeline to move data between different AWS compute and storage services, to schedule external jobs. Hence AWS helped company in delivering highly flexible data-processing service to a worldwide customer base. The company is able to achieve 100% processing speed for its iterative jobs, making its systems faster and more efficient.

Benefits:

● Simple -drag & drop of templates to make visual pipeline on console for various tasks like logfile processing, archiving data, etc
● Reliable - highly fault tolerant, it automatically retries the activity in case of failure
● Flexible - offers features like scheduling, tracking, error handling, executing SQL queries directly against DB, etc
 

c.  Amazon Elastic Map Reduce

Amazon EMR provides elastic infrastructure framework of dynamically scalable Amazon EC2 along with Amazon S3 to provide framework like frameworks such as Apache Hadoop, Apache Spark, and Presto. It helps in data & financial analysis, data warehousing, web indexing, etc.

Benefits: –

● Easy to use – easy to use, set up, configure, etc.
● Reliable – automatically retries failed tasks and replaces instances which are low performing
● Elastic – increase or decrease number of instances at any moment
 

As per Shazam[11] case study in AWS, the company connects people around the world to the music, TV shows and brands they love. With this app people identify a song, a TV program, or ad and get relevant information about it on a single click. The company faced spike in demand on service during their Super Bowl advertising campaign. Soon they switched to cloud technologies and implemented new solution using AWS EC2, EBL, DynamoDB, etc. The company is using Amazon EMR for large-scale data analysis which require more than 1 million writes/sec. Implementing AWS solution enabled company to handle high traffic online in a short period.

Conclusion
 

AWS provide secure & efficient solutions for companies to grow faster in a cost-effective manner by providing a wide set of global compute, storage, database, analytics, application, IoT, etc. services.
Companies have successfully integrated their remote network or Amazon VPC network using AWS resources. With AWS, companies no longer worry about underlying infrastructure or its location but focus on developing their applications.

References
 

• Amazon Web Services. (2018). AWS Named as a Leader in The Forrester Wave: Database-as-a-Service Q2 2017 Available at: https://aws.amazon.com/blogs/database/aws-named-as-a-leader-in-the-forrester-wave-database-as-a-service-q2-2017/
• Armbrust, M. and Fox, A. (2009). Above the Clouds: A Berkeley View of Cloud Computing.
• Docs.aws.amazon.com. (2018). “AWS Documentation” Available at: https://docs.aws.amazon.com/index.html#lang/en_us
• Gartner.com. (2018). Gartner Says Worldwide IaaS Public Cloud Services Market Grew 29.5 Percent in 2017. [online] Available at: https://www.gartner.com/newsroom/id/3884500
• G2techgroup.com (2018). Why AWS? Available at: https://www.g2techgroup.com/why-aws/
• NASA/JPL’s Mars Curiosity Mission Case Study (2018)– Amazon Web Services (AWS). Available at: https://aws.amazon.com/solutions/case-studies/nasa-jpl-curiosity/
• Robitaille, V. (2017).” Understanding AWS Elastic Load Balancing” Available at: https://medium.com/nubego/understanding-aws-elastic-load-balancing-62bf3d2cec5f
• Salesforce UK Blog. (2018). “Why Move to The Cloud? 10 Benefits of Cloud Computing” Available at: https://www.salesforce.com/uk/blog/2015/11/why-move-to-the-cloud-10-benefits-of-cloud-computing.html
• www.tutorialspoint.com.(2018) “Amazon Web Services Tutorial” Available at: https://www.tutorialspoint.com/amazon_web_services/index.htm

[1] Amazon Web Services, Inc. (2018). What is AWS? – Amazon Web Services. [online] Available at: https://aws.amazon.com/what-is-aws/

[2] Airbnb Case Study (2018) – Amazon Web Services (AWS). Available at: https://aws.amazon.com/solutions/case-studies/airbnb/

[3] Wipro Case Study (2018) – Amazon Web Services (AWS) Available at: https://aws.amazon.com/solutions/case-studies/wipro/

[4] Serverless Case Study (2018) — Netflix – DZone Cloud Available at: https://dzone.com/articles/serverless-case-study-netflix  
&  Netflix & AWS Lambda Case Study(2018) – Amazon Web Services (AWS) Available at: https://aws.amazon.com/solutions/case-studies/netflix-and-aws-lambda/

[5] Amazon Aurora Customer Testimonials (2018) – Amazon Web Services (AWS) Available at: https://aws.amazon.com/rds/aurora/customers/

[6] Duolingo Case Study-DynamoDB (2018)-Amazon Web Services (AWS) Available at: https://aws.amazon.com/solutions/case-studies/duolingo-case-study-dynamodb/

[7] AWS Case Study: PlaceIQ. (2018)Available at: https://aws.amazon.com/solutions/case-studies/placeiq/

[8] Zimplistic Case Study (2018) – Amazon Web Services (AWS) Available at: https://aws.amazon.com/solutions/case-studies/zimplistic/

[9] AWS Case Study: BuildFax & Amazon Machine Learning (2018) Available at: https://aws.amazon.com/solutions/case-studies/buildfax-and-amazon-machine-learning/

[10] Salesforce DMP Case Study (2018)-Amazon Web Services (AWS) Available at: https://aws.amazon.com/partners/success/salesforce-case-study/

[11] AWS Case Study: Shazam(2018) Available at: https://aws.amazon.com/solutions/case-studies/shazam/