Serverless Computing: Advantages, Limitations and Use Cases

Vamsi Krishna Thatikonda 

8921 Satterlee Ave Se, Snoqualmie, WA

Suggested Citation Abstract:

Thatikonda, V.K. (2023). Serverless computing, also known simply as "serverless," represents
Serverless Computing: Advantages, a significant transformation in how applications are designed,
Limitations and Use Cases. deployed, and executed. Contrary to traditional methods where
European Journal of Theoretical and developers manage servers, serverless allows them to focus only on
Applied Sciences, 1(5), 341-347. coding, trusting cloud providers with infrastructure responsibilities.
DOI: 10.59324/ejtas.2023.1(5).25 This paradigm shift results in dynamic resource allocation, where
users are billed based on actual resource consumption rather than
pre-allocated capacities. The article elucidates serverless computing's
nature, highlighting its core components - Function as a Service (FaaS) and Backend as a Service (BaaS).
The benefits of serverless are underscored, including cost efficiency, inherent scalability, rapid
development, and reduced operational demands. Yet, it is not without limitations. Concerns such as "cold
starts," potential vendor lock-in, restricted customization, and specific security vulnerabilities are
discussed. Practical serverless applications include web applications, data processing, IoT backends,
chatbots, and ephemeral tasks. In conclusion, while serverless computing heralds a new age in cloud
technology, businesses are encouraged to discerningly evaluate its pros and cons, mainly as the landscape
evolves. The future is serverless, prompting organizations to determine their readiness for this revolution.

Keywords: Serverless Computing, Function as a Service (FaaS), Backend as a Service (BaaS), Scalability, Cold Starts,
and Vendor Lock-in.

Introduction simplified development processes. Traditional

server-based infrastructure requires developers
Serverless computing, often called simply to manage, maintain, and scale servers, often
"serverless," represents a paradigm shift in how leading to inefficiencies in resource use and
applications are developed, deployed, and increased operational costs. The transition to
executed. Instead of focusing on server serverless is a response to these challenges. It
management, developers rely on cloud providers offers a higher level of abstraction, removing the
to handle the infrastructure, allowing them to need to interact with the underlying
concentrate solely on the code (Castro et al., infrastructure directly, expediting development
2019). This computing model dynamically cycles, and reducing overheads (Baldini et al.,
manages the allocation of machine resources, 2017).
billing users based on the actual number of
resources consumed by executions rather than
on pre-purchased capacity units (Castro et al., What is Serverless Computing?
2019).
Despite its name, serverless computing does not
Consistent efforts to abstract have marked the imply the absence of servers. Instead, it
evolution of computing paradigms and

This work is licensed under a Creative Commons Attribution 4.0 International License. The license permits unrestricted
use, distribution, and reproduction in any medium, on the condition that users give exact credit to the original author(s)
and the source, provide a link to the Creative Commons license, and indicate if they made any changes.
emphasizes the developer's detachment from Traditional computing models necessitated
server management, allowing them to focus intricate involvement in infrastructure setup,
exclusively on the application's functionality scaling, maintenance, and updates. On the other
(Shafiei, Khonsari, & Mousavi, 2022). This hand, serverless abstracts these complexities,
approach relies on cloud service providers to offering developers an environment where the
dynamically allocate and execute backend code, underlying infrastructure becomes invisible,
billing users based on actual compute processes handled entirely by the cloud provider (Shafiei,
rather than predetermined server capacity Khonsari, & Mousavi, 2022).
(Shafiei, Khonsari, & Mousavi, 2022).

Figure 1. Traditional vs Serverless Computing Schemes

Two principal components of serverless go model ensures businesses are billed based on
architectures are Function as a Service (FaaS) actual resource consumption rather than
and Backend as a Service (BaaS). FaaS involves allocated server capacity (Schleier-Smith et al.,
executing code snippets in response to events, 2021). This approach eliminates the expenditure
eliminating the need for long-running server tied to idle server time, allowing companies to
processes. Examples include AWS Lambda and optimize costs in real-time. Furthermore, the
Azure Functions (Eismann et al., 2021). absence of a dedicated infrastructure to manage
Conversely, BaaS provides ready-to-use backend translates to a significant decrease in associated
services, such as databases or authentication overheads. Organizations adopting serverless
processes, further simplifying application reported a 60% reduction in infrastructure
development (Eismann et al., 2021). management costs compared to traditional cloud
models (Li et al., 2021). Further, the one-time
deployment cost of Serverless computing is 68%
Advantages of Serverless Computing less than server-based computing, as seen in the
following table.
Cost Efficiency
One of the most enticing attributes of serverless
computing is its cost efficiency. The pay-as-you-

342

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5

Scalability beneficial for applications that experience
fluctuating traffic loads. For instance, e-
Serverless computing offers inherent scalability.
commerce sites during sale events or startups
Cloud providers automatically handle scaling
experiencing rapid user growth can leverage
based on the application's demand, ensuring
serverless to meet dynamic demands without
seamless operation without manual intervention
incurring the costs and challenges of pre-scaling.
(Cloudflare, n.a.). This feature is particularly

Table 1. Comparison of Deployment Cost

One-time Development Cost for Server-based (EC2) Compared with Serverless (QWS lambda)
Development Server-based Serverless Difference
Days to Deploy ~25 days ~8 days ~17 days
One-time Upfront $38,300 $12,300 $26,000
Monthly Cost $640 $205 $(435)
Serverless is 68% cheaper
than server-based
Source: Arora, Tayal, & Sembhi, (2021)

Figure 1. Benefits of Cost Savings in Dynamic Scaling

Source: Arora, Tayal, & Sembhi, (2021)

Improved Development Speed Reduced Operational Overheads

Serverless architectures speed up software The operational demands in traditional
development (Rosenbaum, 2017). Without the computing models, ranging from server
need to manage infrastructure, developers can management to software patching, often divert
focus solely on writing and refining code, valuable time and resources from core business
drastically reducing development cycles (Castro objectives. By outsourcing these tasks to cloud
et al., 2019). Additionally, with the immediate providers, serverless computing essentially
execution environment offered by serverless eradicates these overheads (Castro et al., 2019).
platforms, rapid deployment, and iteration Deployment becomes a straightforward process,
become the norm. This quick turnaround means devoid of the complexities tied to server
businesses can respond faster to market changes, configurations. Moreover, rollback, reverting to
user feedback, and emerging trends. previous versions of applications in case of
errors or issues, is vastly simplified (Lloyd, 2022).
This ensures businesses can maintain continuity
in face of challenges.

343

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5

Limitations and Concerns of Vendor Lock-in
Serverless Computing Adopting serverless architectures often means
While serverless computing offers numerous aligning with a specific provider's toolset,
advantages, it's essential to understand the infrastructure, and services. This dependence
associated challenges and concerns. can lead to vendor lock-in, where migrating to a
different platform becomes time-consuming and
Cold Starts costly due to incompatible interfaces and
One of the most discussed limitations in configurations (Taibi, Spillner, & Wawruch,
serverless computing is the phenomenon of cold 2021). This lock-in can limit flexibility, making it
starts. A cold start occurs when a new instance challenging for organizations to adapt to
of a function is initiated, and there's a latency evolving business needs or leverage better
involved before the Function begins executing, offerings in the market.
primarily because the environment needs to be Limited Customization and Control
set up from scratch (Vahidinia, Farahani, &
Aliee, 2020). This latency can affect As serverless platforms manage the underlying
performance, particularly in applications where infrastructure, they often impose certain
swift response times are critical. For user-facing restrictions on runtime environments, memory
applications, the additional latency introduced limits, and execution durations. This can limit
by cold starts can result in sub-optimal user the scope for deep customization, which might
experiences, potentially impacting user retention be essential for some niche applications. While
and engagement. simplifying many operations, serverless
frameworks may not cater to highly specialized
use cases, necessitating workarounds or
alternative solutions (Baldini et al., 2017).

Figure 2. Use Case of Serverless Computing

Source: Adam, J. (2022)

Security Concerns security remains a shared responsibility. This

model means developers are responsible for
With serverless, while the provider primarily
securing their application code, configurations,
handles infrastructure security, the application's

344

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5

and third-party libraries (Marin, Perino, & Di IoT Backend
Pietro, 2022). Serverless environments may also The Internet of Things (IoT) is characterized by
introduce unique security vulnerabilities. For
sporadic data bursts from numerous devices.
instance, the dynamic nature of serverless may
Managing a consistent server setup for such
expose applications to event-data injections or
erratic data inflows can be cumbersome.
potential unauthorized function invocations
Serverless platforms shine here, seamlessly
(Marin, Perino, & Di Pietro, 2022). Functions,
handling sudden data influxes and ensuring
especially those interacting with other cloud
efficient processing and storage without manual
services, can become possible entry points for
intervention (Cicconetti, Conti, & Passarella,
attackers, making a solid case for rigorous
2021).
security practices in serverless architectures
(O'Meara, & Lennon, 2020). Chatbots and AI Services
In customer service and support, AI-powered
chatbots have gained immense traction.
Use Cases for Serverless Computing Serverless provides an ideal environment for
The use cases for serverless computing are these chatbots, ensuring quick responses by
numerous; however, the four prominent use eliminating the need for maintaining a complete
cases are web applications, data processing, IoT backend (Lehvä, Mäkitalo, & Mikkonen, 2018).
Backends, Chatbots and Services, and The inherent scalability of serverless ensures that
Ephemeral tasks. these bots can cater to many users
simultaneously, offering consistent
Web Applications performance.
Serverless computing has proven instrumental in Ephemeral Tasks
the modern web development. Web
applications, particularly those requiring Intermittent Jobs, such as CRON jobs or other
dynamic content generation, can leverage scheduled tasks, can be inefficient on traditional
serverless functions to serve content in real-time servers due to resource wastage during idle
without server management overhead (Castro et times. With serverless, these tasks can be
al., 2019). This approach offers scalability, executed on-demand, ensuring optimal resource
ensuring the application remains responsive usage and cost efficiency (Lynn et al., 2017).
during traffic surges. Leveraging serverless,
developers can focus solely on front-end
development, knowing that the serverless Conclusion
provider optimizes and maintains the backend With its dynamic scalability and cost-efficiency,
processes. serverless computing marks a transformative
Data Processing shift in cloud technology. While the advantages
are compelling, businesses must also weigh the
The sheer volume and velocity of data demand challenges, particularly in customization and
efficient processing solutions in today's digital potential vendor lock-in. As technological
world. Serverless frameworks cater to this need innovations advance, serverless is poised to
by enabling data transformation, analysis, and evolve, potentially addressing its current
processing (Goli et al., 2020). As data streams in, limitations. The coming years might see a
serverless functions are invoked to process and convergence of serverless with other
analyze the data instantaneously, making it technologies, ushering in a new age of integrated
especially beneficial for applications requiring digital solutions. The call is clear for businesses
immediate insights. For instance, serverless eyeing the future: delve deeper, critically assess,
computing systems can be applied in the and determine if serverless aligns with your
financial sector to fasten data processes while strategic objectives and technological
reducing costs (Goli et al., 2020).

345

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5

aspirations. The future is serverless; the question (2020). Migrating from monolithic to serverless: A
remains, are you ready for it? fintech case study. Companion of the ACM/SPEC
International Conference on Performance
Engineering.
References https://doi.org/10.1145/3375555.3384380
Adam, J. (2022). Serverless architecture for Lee, H., Satyam, K. & Fox, G. (2018). Evaluation
increased Dev Productivity. K&C. Retrieved of production serverless computing environments. In 2018
from https://kruschecompany.com/serverless- IEEE 11th International Conference on Cloud
architecture-for-modern-apps-providers-and- Computing (CLOUD).
caveats/ https://doi.org/10.1109/cloud.2018.00062
Arora, G., Tayal, A. & Sembhi, R. (2021). Lehvä, J., Mäkitalo, N., & Mikkonen, T. (2018).
Determining the Total Cost of Ownership: Case study: Building a serverless messenger
Comparing Serverless and Server-based chatbot. Current Trends in Web Engineering, 75–86.
Technologies. Retrieved from https://doi.org/10.1007/978-3-319-74433-9_6
https://www.strategist- Li, Z., Tan, Y., Li, B., Zhang, J. & Wang, X.
hub.com/pdfs/203755_AWS_MAD_TCO_eB (2021). A survey of cost optimization in
ook_2021_Final.pdf serverless cloud computing. Journal of Physics:
Baldini, I., Castro, P., Chang, K., Cheng, P., Conference Series, 1802(3), 032070.
Fink, S., Isahagian, V., … & Suter, P. (2017). https://doi.org/10.1088/1742-
Serverless Computing: Current Trends and 6596/1802/3/032070
Open Problems. In: Chaudhary, S., Somani, G., Lloyd, J. (2022). Containers and Serverless. In
Buyya, R. (eds) Research Advances in Cloud Infrastructure Leader's Guide to Google Cloud: Lead
Computing. Springer, Singapore. Your Organization's Google Cloud Adoption, migration
https://doi.org/10.1007/978-981-10-5026-8_1 and modernization journey. New York: Apress.
Castro, P., Ishakian, V., Muthusamy, V., & Lynn, T., Rosati, P., Lejeune, A. & Emeakaroha,
Slominski, A. (2019). The rise of serverless V. (2017). A preliminary review of Enterprise Serverless
computing. Communications of the ACM, 62(12), Cloud Computing (Function-as-a-service) platforms. In
44–54. https://doi.org/10.1145/3368454 2017 IEEE International Conference on Cloud
Cicconetti, C., Conti, M. & Passarella, A. (2021). Computing Technology and Science
A decentralized framework for serverless edge (CloudCom).
computing in the internet of things. IEEE https://doi.org/10.1109/cloudcom.2017.15
Transactions on Network and Service Management, Marin, E., Perino, D. & Di Pietro, R. (2022).
18(2) 2166–2180. Serverless Computing: A security perspective.
https://doi.org/10.1109/tnsm.2020.3023305 Journal of Cloud Computing, 11(1).
Cloudflare. (n.a.). Why use serverless https://doi.org/10.1186/s13677-022-00347-w
computing? Pros and cons of serverless. O'Meara, W. & Lennon, R.G. (2020). Serverless
Retrieved from Computing Security: Protecting Application logic. In
https://www.cloudflare.com/learning/serverles 2020 31st Irish Signals and Systems Conference
s/why-use-serverless/ (ISSC).
Eismann, S., Scheuner, J., Eyk, E., Schwinger, https://doi.org/10.1109/issc49989.2020.91802
M., Grohmann, J., Abad, C., & Iosup, A. (2021). 14
Serverless applications: Why, when, and how?," Rosenbaum, S. (2017). Serverless Computing in
IEEE Software, 38(1), 32–39. Azure with. NET. Packt Publishing Ltd..
https://doi.org/10.1109/ms.2020.3023302
Schleier-Smith, J., Sreekanti, V., Khandelwal, A.,
Goli, A., Hajihassani, O., Khazaei, H., Carreira,J., Yadwadkar, N.J., Popa, R.A.,
Ardakanian, O., Rashidi, M. & Dauphinee, T. Gonzalez, J.E., Stoica,I., & Patterson, D.A.

346

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5

(2021). What serverless computing is and should Vahidinia, P., Farahani, B. & Aliee, F.S. (2020).
become: The next phase of cloud computing. Cold start in Serverless Computing: Current trends and
Communications of the ACM, 64(5), 76–84. mitigation strategies. In 2020 International
https://doi.org/10.1145/3406011 Conference on Omni-layer Intelligent Systems
(COINS).
Shafiei, H., Khonsari, A. & Mousavi, P. (2022).
https://doi.org/10.1109/coins49042.2020.9191
Serverless Computing: A Survey of
377
Opportunities, Challenges, and Applications.
ACM Computering Survey, 54(11s), 239. Weston, R. (n.a.). Serverless architectures and
https://doi.org/10.1145/3510611https://doi.o continuous delivery. GoCD. Retrieved from
rg/10.31224/osf.io/u8xth https://www.gocd.org/2017/06/26/serverless-
architecture-continuous-delivery/
Taibi, D., Spillner, J. & Wawruch, K. (2021).
Serverless Computing-Where are we now, and
where are we heading? IEEE Software, 38(1), 25–
31. https://doi.org/10.1109/ms.2020.3028708

347

www.ejtas.com EJTAS 2023 | Volume 1 | Number 5