Serverless is driving passionate debate in social media just like how public cloud ramped up the discussions circa 2008. We are once again seeing patterns in discussions that mirror what we saw in the early days of cloud. The increasing pitch of some of these messages (by the respective advocates or critics) are not helping anyone. In this post, we want to debunk some of these claims and help set the tone.
Serverless is the magic pill to cure all ill
One of the underlying tones in the serverless advocacy is that it is the magic pill for all the IT problems. In fact, I saw one slide from the recently concluded ServerlessConf where the speaker was making fun of Kubernetes users. Yes, serverless is the highest level of developer platform abstraction that does all the heavy lifting needed to deploy the functions. But it is still severely opinionated and not suitable for all kinds of workloads. In fact, it is even limited for some of the stateless workloads. Dismissing any technology other than serverless as unnecessary heavy lifting is simplistic. Yes, if you invest heavily in running your own data centers, it is an unnecessary heavy lifting. Even here, it is not a binary characterization. What developers will be using, in the foreseeable future, are a continuum of services starting with managed Kubernetes services to serverless containers to serverless functions. Any claims about serverless as the only service that matters is either simplistic thinking or just pure hallucination.
The false dichotomy of public transport and driving your own cars
On the other side of the debate, we have people making claims that Kubernetes is the future and serverless is just a crazy talk by a bunch of developers. In fact, the “Kubernetes is the only future” camp was involved in a conversation with someone advocating serverless on Twitter. The counter-argument against serverless, especially Functions as a Service offering like AWS Lambda was as follows:
If people should use serverless to avoid the heavy lifting, then everyone should be using public transport than driving their own car
Even though I personally believe that using public transport is the right thing to do to avoid climate change, I find this comparison too simplistic for a rebuttal. Public transport, at least in the USA, is not widespread enough to be an alternative agile mode of transportation for driving your own car. But, in the case of public cloud services, there are no such bottlenecks. The consumption of a service like AWS Lambda is as simple as opening the web browser and signing up with one’s credit card.
Let us not do advocacy from the extremes
Yeah, this post is more of a rant than a typical StackSense post but it is time we do not take a binary position on any technology. That is an easy way to make any curious enterprise user drag their feet on any new technology. Instead, let us engage in discussions that actually helps enterprise users understand the landscape better and embrace modern technologies.