Why Serverless Computing Can Create A Revolution In Cloud Technology
Parvathy Prakash Akshay Jeevan
PG Student PG Student
Department of Computer Applications (MCA) Department of Computer Applications (MCA)
Mar Athanasius College of Engineering Mar Athanasius College of Engineering
Kothamangalam, Kerala, India Kothamangalam, Kerala, India
[email protected] [email protected]Ajay Krishnan Nisha Markose
PG Student Assistant Professor
Department of Computer Applications (MCA) Department of Computer Applications (MCA)
Mar Athanasius College of Engineering Mar Athanasius College of Engineering
Kothamangalam, Kerala, India Kothamangalam, Kerala, India
[email protected] [email protected]ABSTRACT
Serverless computing is a cloud-computing execution 1. INTRODUCTION
model in which the cloud provider runs the server, Cloud computing is the delivery of computing services
and dynamically manages the allocation of machine servers, storage, databases, networking, software,
resources. Organizations that are keen to leverage analytics, and intelligence over the Internet (“the cloud”)
modern technology to gain a disruptive edge, optimal to offer faster innovation, flexible resources, and
efficiency, advanced agility and save cost are adopting economies of scale. Consumers pay only for cloud
these architectural styles rapidly. A company that gets services they use, run their infrastructure more efficiently
backend services from a serverless vendor is charged and scale as their business needs change. In serverless,
based on their computation and do not have to reserve developers don't need to setup and administer the servers
and pay for a fixed amount of bandwidth or number running back-end applications. The back-end
of servers, as the service is auto-scaling. Note that infrastructure is maintained by third party providers and
although called serverless, physical servers are still the functionality needed is offered as services.
used but developers do not need to be aware of them. Using Serverless computing, developers can develop
This review paper presents a comprehensive study on applications at rapid pace with reduced costs without
serverless computing architecture and also extends an worrying about infrastructure. The servers still exist in
experimentation of the working principle of serverless Serverless computing but managed dynamically based on
computing reference model adapted by AWS demand by Cloud service provider instead of the
Lambda. This paper also exploring the advantages application owner.
and limitations. The various research avenues in Function as a Service (FaaS) is a small discrete reusable
serverless computing are identified and presented. chunk of code which is stateless compute container
Keywords - cloud computing, serverless computing, modelled for an event-driven solution. User can simply
FaaS, AWS Lambda, Microsoft Azure Function, upload modular chunks of functionality into the cloud that
Google Cloud Function are executed independently. Function as a Service (FaaS)
1
is basically a combination of self-managed functions and 2.2 DISADVANTAGES OF PAAS
cloud services that serve as the Serverless application. 2.2.1 DEPENDANCY ON VENDOR
The event or trigger-based computing helps to cut cost Your business will be governed by the provider’s
and enables more efficient application development since functional capabilities, speed and reliability. That’s why
additional resources are provisioned only when a pre- it is very important to be confident in your business
defined event occurs which ensures resources are not idle. partner and be prepared for some unforeseen
The allocation of resource on cloud is managed by service circumstances that may happen to even the most reliable
provider dynamically and billing is based on the actual provider. At the very least, you should perform your own
resources consumed by a running application. data backup.
2.2.2 SECURITY RISKS
2. WHAT IS PAAS? PaaS software is accessible in a public environment
Platform as a Service (PaaS) represents a comprehensive where multiple end users have access to exactly similar
cloud- based runtime environment with resources that resources. For some applications that contain sensitive
allow customers to create both simple and advanced apps. data, this is not a good option.
It works like providers deliver pay-as-you-go resources • MIXED UP COMPLEXITIES
and support secure network connections, while clients Some of the applications developed could also be local
manage programs and services they create on the platform. while others are from the cloud which may increase the
The PaaS service is used for the support of a full lifecycle complexity.
of a web application:
• Development 3. SERVERLESS COMPUTING
• Testing Serverless computing allows us to build and run
• Deployment applications and services without worrying about server
• Management provisioning, scaling and how to manage it. The
• Updating infrastructure can scale to any extent with manual or auto
provisioning of new servers on sudden spike in usage. It
2.1 PAAS CONSISTS OF :
saves development time and helps developers to focus on
• Physical infrastructure — data centers, servers, their core business logic coding instead of worrying about
storage, network equipment (Components of managing and operating servers or runtimes.
IaaS).
• Intermediate-layer software — operating
4. SERVERLESS COMPUTING A
systems, libraries, languages, frameworks, tools
REVOLUTION IN CLOUD
for building apps.
• User interface — a graphical user interface COMPUTING
(GUI), a command line interface (CLI) or both, • IT SAVES TIME AND REDUCE
the interface displays the architecture and logic of OVERHEAD COSTS
the deployment machinery, providing usability One of serverless computing top advantages is its ‘pay-
and learnability of a given PaaS. as-you-use’ model, since you pay for the time your code
is executed and the number of times it has been triggered.
We have analysed that as serverless functions are short-
2
lived and are automatically deallocated by the provider, attacks prevention, and many other security aspects are
FaaS reduces the risk of incurring unwanted cloud costs, quite solid in serverless computing.
which is a significant plus in start-ups and small • QUICK DEPLOYMENTS AND UPDATES
companies. FaaS allows developers to run code on any ARE POSSIBLE
type of application or backend service without Using a serverless infrastructure, there is no need to
provisioning or managing servers and uploads it to a upload code to servers or do any backend configuration in
provider. That means ‘takes care of all hardware order to release a working version of an application.
provisioning, virtual machine and container management, Developers can very quickly upload bits of code and
and even tasks like multithreading that often are built into release a new product. They can upload code all at once
application code’. or one function at a time, since the application is not a
• STOP WORRYING ABOUT single monolithic stack but rather a collection of functions
INFRASTRUCTURE PROBLEMS provisioned by the vendor. This also makes it possible to
With serverless computing, you do not have to worry quickly update, patch, fix, or add new features to an
about hardware capacity requirements or server application.
reservations. Developers can stop worrying about how • CODE CAN RUN CLOSER TO THE END
many machines are needed at any given point of time, USER, DECREASING LATENCY
especially during peak hours, whether the machines are Because the application is not hosted on an origin server,
working optimally, whether all the security measures are its code can be run from anywhere. It is therefore possible,
offered and so on. depending on the vendor used, to run application
• FUNCTION AS A SERVICE (FAAS) functions on servers that are close to the end user. This
Function as a service (FaaS), similar tasks are grouped reduces latency because requests from the user no longer
together and their business logic is written into small have to travel all the way to an origin server.
pieces of code called functions. when a pre-defined event
occurs it triggers the execution of the function. 4.1 SERVERLESS ARCHITECTURE
• SCALING AS PER THE SIZE OF THE
WORKLOAD
Serverless Computing automatically scales your
application. With each individual trigger, your code will
run parallel to it, thereby reducing your workload and
saving time in the process. When the code is not running,
you don’t have to pay anything. The charging takes place
• Api Gateways — A HTTP server where routes
for every 100ms your code executes and for the number
and endpoints are defined in configuration, and
of times the code is triggered. This is a good thing because
each route is associated with a resource (functions
you no longer pay for an idle compute.
in this case) to handle that route.
• BETTER SECURITY
• The application is separated to 2 different
As seen on Medium, serverless provides well-conditioned
functions that will only be triggered when there’s
security means, as programming flow protection, risks
a request coming in.
mitigation, fraud detection, HTTP floods monitoring,
cross-site scripting forecast, bots identification, DDoS
3
• Our authentication function has been outsourced tasks. Payment is calculated by the 100’s of milliseconds
to a 3rd party service. We no longer have to write of use and total cost of ownership is also better for high-
our own code. load scenarios.
Functions allow a new way to modularise our code giving • GET ALL THE BENEFITS OF
us the flexibility of porting the same functions for a new ROBURST CLOUD
project without changing anything. INFRASTRUCTURE
FaaS offers inherent high availability because it is spread
5. WHAT IS FAAS? across multiple availability zones per geographic region
FaaS (Function-as-a-Service) is a type of cloud- and can be deployed across any number of regions
computing service that allows you to execute code in without incremental costs.
response to events without dealing with the complicated
server setup infrastructure typically involved in cloud-
hosted applications.
For running an application on the internet, you need to set
up and managing a virtual or physical server and
managing an operating system and server software for
hosting processes. With FaaS, the physical hardware,
virtual machine operating system, and web server
6. USE CASES
software management are all handled automatically by
• AUTO-SCALING WEB SITES AND
your cloud service provider. This allows you to focus
APIS
exclusively on individual functions in your application
Websites and applications that use serverless architecture
code not on infrastructure.
doesn’t require a fixed quantity of resources. The backend
of a serverless application is scalable, so that it can handle
5.1 BENEFITS OF FAAS
a sudden surge of traffic.
• FOCUS MORE ON CODE, NOT
• IMAGE AND VIDEO MANIPULATION
INFRASTRUCTURE
It allows you to make performance enhancing images or
Faas allows developers to focus on their actual
video services for applications. It dynamically transforms
application not the infrastructure. Applications are
the image for different target devices.
divided into multiple functions that trigger other services.
With Serverless, you can access the Google Vision API
This allows applications to scale automatically when
or Amazon Rekognition to identify faces and images, or
needed without having to manage the infrastructure. All
flag inappropriate content. You can automatically process
the maintenance activities are done by the service
and reformat user uploaded images, or resize them into
providers like aws lambda, IBM OpenWhisk etc.
specific thumbnail dimensions.
• PAY ONLY FOR THE RESOURCES
• PROCESSING EVENTS AND SAAS
YOU USE, WHEN YOU USE THEM
Modern applications use software from 3rd party SaaS
With FaaS, you pay only when an action occurs. When
providers to deliver functionality. They don’t let you run
the action is done, everything stops—no code runs, no
functions on their system, but they produce events.
server idles, no costs are incurred. FaaS is, therefore, cost-
effective, especially for dynamic workloads or scheduled
4
Serverless lets you assign function to those events, no consumption (e.g.
matter where the functions are hosted. You can use this to throughput,
take action. memory)
• EVENT TRIGGERED COMPUTING Deployment Necessary to make Quick deployments
In multimedia processing business applications, huge and updates changes to the and updates are
volumes of files are frequently uploaded to Object whole application possible by upload
Storage Services [OSS] for processing. The requirements for adding new bits of code and
may be such as transcoding, watermarking, fetching the feature release a new
data. This business scenario involves a variety of devices product.
like desktop computers or PDAs or mobile phones Focus Focus on Focus on
accessing different file types of uploading multimedia infrastructure. application.
content such as images, videos, and text files. Event-
triggered computing will be a solution for addressing 8. CONCLUSION
many technical difficulties by event triggered computing. Serverless computing is an increasingly popular paradigm
that takes this desire to its logical conclusion: With
7. COMPARISON BETWEEN serverless computing, you don’t have to know anything
TRADITIONAL AND about the hardware or OS your code runs on, as it’s all
taken care of for you by a service provider. Naturally,
SERVERLESS COMPUTING
there are still servers involved, but their provisioning and
maintenance are entirely taken care of by the provider.
ACTIVITY TRADITIONAL SERVERLESS This includes nothing that would be bare metal, nothing
CLOUD CLOUD that’s virtual, nothing that’s a container—anything that
COMPUTING COMPUTING involves you managing a host, patching a host, or dealing
Cost Pay for server Pay for consistent with anything on an operating system level, is not
units. throughput or something you should have to do in the serverless world.
execution duration
Security It is in a public It provides security 8.1 FUTURE SCOPE
environment and measures like The serverless computing outlook is connected with lots
resources are programming flow of factors. As long as consumers demand Serverless
shared. protection, risks pattern, there will be providers ready to harvest financial
mitigation, fraud benefit out of it. Even though Prominent Programming
detection etc. languages are supported currently, it should extend
Server Continuous Automatically support to almost all programming languages available.
management management managed by service For this reason, serverless architecture should be able to
required. providers. adapt the everchanging landscape of software
Scaling Automatic scaling, Automatic scaling, development on a perpetual basis. This will leverage the
based on units of through toggling the use of serverless computing towards next level. With the
individual servers. units of present evolution of serverless computing, it seems to be
becoming a safe and agile service that has a bright future.
5
9. ACKNOWLDGEMENT
We would like to thank Prof. Nisha of Dept. of Computer
Applications Mar Athanasius College of Engineering,
Kothamangalam and all other teaching staffs for their
useful advice and feedback.
REFERENCE
[1] Mohit Sewak,Sachchidanad Singh“Winning in the era
of Serverless Computing and Function as a Service”
[2] Dr. R. Arokia Paul Rajan
“Serverless Architecture - A Revolution in Cloud
Computing”
[3] Evolution of Serverless computing, Accessed on 10th
August 2018, https://www.ca.com/us/modern-software-
factory/content/the-evolutionof-serverless-
computing.html
[4] Iron, Accessed on 06th September 2018,
https://www.iron.io/docs/Whitepaper_Serverless_Final_
V2.pdf
[5] Reasons Why Serverless Computing will Create a
Revolution in Cloud Technology,
https://datafloq.com/read/7-reasons-serverless-
computing-revolution-cloud/2871
[6] AWS Lambda
https://aws.amazon.com/lambda/
[7] Why use Serverless Computing? | Pros and Cons
of Serverless,
https://www.cloudflare.com/learning/serverless/why-
use-serverless/
