There is increasing discussion in the IT world about serverless computing, so it is worth learning more about the possibilities of this solution. What are the pros and cons of serverless architecture? When is it worth switching from server to serverless computing, and when is it better not to do so? We're answering these questions below.
Before you choose between server vs serverless computing, you need to know all the pros and cons of these solutions and match them to the nature of your business.
<blockquote><p>"Since its migration [to serverless architecture], Coca-Cola has reduced its operational costs by 40 percent and IT ticket volume by 80 percent"</p><p>source: AWS Coca-Cola case study</p></blockquote>
Operational cost reduced by 40%? Ticket volume lowered by 80%? Sounds like a wet dream of virtually every technology-based business.
Such case studies lead us to believe that serverless is the solution to most of our software engineering pains.
But it's NOT.
In this article, we're exploring the hidden costs of serverless computing, and help you decide if it's the right fit for your business.
Serverless computing comes at a cost — both monetary and non-monetary. Here's what you need to remeber when considering the switch to cloud architecture.
When you choose serverless computing, you don't manage the server. Thus, you lose control of the server hardware, execution environments and updates. These may become a barrier for certain products.
Some serverless developers claim that processing times can vary widely between runs, likely due the code may be deployed on a different server with different specifications when restarted.
A known performance problem on some serverless platforms is also Cold Start. This is the time required to call a new container with a specific function if no instance is currently active.
Serverless computing means no dedicated server at your disposal. Serverless providers can therefore run code for several clients on a single server.
This "multi-tenancy" can negatively affect the security and performance of the application in case of multiple tenant access or incorrect configuration. The result can be data leaks or greater vulnerability to hackers.
Serverless architectures are billed for code runtime. Therefore, operating applications that require long processes in a serverless infrastructure may end up more costly than traditional infrastructure.
Each serverless provider offers a unique set of features. Therefore, using the entire range of backend services from a particular provider makes the relationship between the customer and the vendor long-term. In addition, you should know that configuring a serverless architecture with a specific supplier can make migration challenging.
With traditional application models, developers often install versions of software components locally to test whether the code works before it is uploaded to the production environment.
Serverless applications may consist of many separate features, therefore replicating the production environment may not be possible.
Debugging also becomes more complicated because developers have no visibility into backend processes. This is also influenced by the fact that the application is divided into many smaller functions.
Switching to serverless computing will be a BAD choice if:
Be wary however, as sticking to traditional architecture and skipping the serverless requires:
Every application implementation model has its flaws. But they need to be weighed against the conveniences a specific technology provides.
In the case of serverless application deployment, these advantages include:
Serverless infrustructure delegates the management of resources to a cloud service provider and allows engineering teams to focus on the business logic of their application instead of dealing with hardware problems.
Switching to serverless computing will be a GOOD choice if:
The products described above will undoubtedly reap all of the benefits of choosing serverless architecture. However, there also are more nuanced cases of projects that could move to serverless with a success.
Switching to serverless computing is worth considering also when you:
Serverless computing owes its popularization primarily to the almost universal adoption of cloud computing and microservices. These, combined with the growing demand for more significant innovation, scalability, and agility without significantly increasing costs, drives users toward serverless solutions. That's why we see more and more organizations switiching to serverless infrastructure.
Companies that moved to serverless architecture with great success include:
Cloud computing is a modern IT technology for processing and storing data on servers outside the local network in a virtual space. Connecting to Cloud Computing means using an Internet connection to access computing resources hosted in data centers managed by third-party providers instead of owning them and keeping them locally.
There are three types of cloud computing:
With the help of cloud computing, developers can run applications in modern data centers with more automation. However, many cloud computing solutions still require a high extent of infrastructure management. Serverless architecture is the subsidiary of cloud computing that eliminates the management part and allows developers to focus on code and business logic.
Let's now examine what differentiates serverless architecture from servers:
Servers are not seamlessly scalable — they have a predetermined storage capacity that is sometimes difficult to scale up, and their resources remain available even when not in use. On the other hand, serverless solutions are fully scalable and allow you to use the computing power and storage you need at any given time.
Unlike a serverless architecture, a server needs to be set up, maintained, and managed.
Server usage fees are billed upfront for a defined service. Serverless systems are billed per function call.
According to Microsoft, Platform as a Service (PaaS) is a complete development and deployment environment in the cloud, including resources to deliver any solution. In this case, the infrastructure is no longer our concern, as the vendor is responsible for providing and maintaining it.
The scope practically coincides with the capabilities of serverless architecture. However, the difference can be seen in scaling — in the case of serverless, scaling works automatically. In the case of PaaS, the user has to configure the scaling parameters himself.
Moreover, unlike serverless computing, PaaS does not offer the possibility to adjust prices depending on usage dynamically. Instead, it provides more control, making testing and debugging applications in a PaaS environment easier.
Serverless computing is a cloud services model in which the developer focuses solely on creating the business logic and not on the infrastructure on which it is to be executed.
Setting up machines, updating operating systems, configuring networks, and scaling applications are all tasks of a given serverless service provider — for example, AWS Lambda, Azure Functions, IBM OpenWhisk or Cloud Functions from Google Cloud.
Serverless computing is a model in which the cloud provider is responsible for executing a piece of code by dynamically allocating resources. Unlike the traditional model, where you buy specific resources and pay for them even if they are not fully deployed, serverless computing provides the option of paying for the resources used.
The term "serverless" is somewhat misleading because it implies the absence of servers. Of course, servers do exist, but developers don't have to deal with them. As a result, they can entirely focus on developing applications instead of worrying about configuring, scaling and securing virtual machines, servers or containers.
Serverless computing is sometimes referred to as "Functions as a Service".
Key advantages of serverless computing include:
Become a better tech leader.
Join 200+ CTOs, founders and engineering managers and get weekly bite-sized leadership lessons that take <60 seconds to read.
No previous chapters
No next chapters