What is Serverless Computing?
Serverless computing isn’t actually serverless. It doesn’t mean the elimination of servers from applications and IT architecture. Serverless is a method of programming for the cloud that changes the way you build and deploy applications. Another way of looking at it: serverless computing makes it easier to write code that has a well-defined function and that operates on a cloud-based platform.
Benefits of Serverless Computing
Of prime importance for CEOs, CFOs or CIOs wanting effective business IT support, serverless computing means improved IT efficiency and, by extension, saving money. Here are three obvious cost-saving benefits of serverless computing.
Effective Utilization
In the cloud model historically used by Amazon Web Services (AWS), virtual machines (VMs) or bare-metal servers are self-contained units. With the new AWS Lambda serverless model, IT service involves leasing a function, essentially a code that does a specific job. Also, you only lease the code for the time in which it is operating.
Better IT Security
When a developer is constrained to developing code that will function in a serverless context, the code will have greater levels of security meaning that data protection is GDPR compliant.
Quicker Turnaround
This is the real good news – the time to production from developing code to deploying it – is significantly reduced. Optimistically, from days to months. This, in itself, saves a bundle.
The Downside
Of course, it can’t all be good news. Serverless computing can impact SLAs, have execution limitations, and DevOps problems.
SLAs Issues
Service level agreements still need to be established for serverless computing and Function-as-a-Service (FaaS). Performance indicators for FaaS functions are so unclear that it is difficult for providers like AWS to make any promises.
Execution Limitations
AWS Lambada currently limits the number of concurrent applications that you can run on the platform. This means that if somebody is doing a load test and running 500 FaaS executions, then you may experience DoS of production applications. This means IT downtime.
DevOps Problems
With serverless computing, the development is alleviated of the responsibility of understanding the requirements of the system hosting the code. To achieve the goals of DevOps and for full IT functionality, developers and operators need to know and understand what each is doing.
Last Words
According to the RightScale 2018 State of the Cloud report, serverless computing is growing at a rate of 75%, an achievement way ahead of its peers. In a recent survey of benefits versus challenges of FaaS, cost reduction came in first as the biggest reduction. The biggest challenge: the ability to monitor and debug code.
Remember though, that serverless computing almost has limitless scalability. You can scale as much as you like at a fraction of the cost since you only pay when a function is used and you are not paying for unused server time. Makes you think!
