Serverless Influencers Make Predictions - Winners, Losers and Mainstream Adoption

The faces of serverless world: (from upper left): David Wells, Yan Cui, Adnan Rahic, Elliot Forbes, Rupak Ganguly, Raymond Camden, Tom McLaughlin, John McKim.

Serverless community is growing bigger and stronger with every week and although this field is fairly new, I have already noticed some serverless superstars popping up — people who are educating, talking, writing and tweeting all things serverless.

These are the people leading the serverless revolution and I figured that their opinions and thoughts would give priceless insights into where it’s all going. So I contacted some enthusiastic influencers and asked their opinion about the future of serverless. Here’s what they predicted…

This article is showcasing only some of the predictions these influencers shared with me, read the full interviews in unedited form here!

Bold predictions — where is “serverless” by the end of 2018?

More than 70% of all newly founded startups will use serverless architecture in some form or another in their product. (Adnan Rahic)

Rise of startups focusing on building tooling around managing functions, team collaboration, observability, monitoring, distributed tracing, debugging, security, logging, and auditing. (Rupak Ganguly)

We’ll agree that serverless does in fact use servers and we’ll move onto bigger and more important discussions. I expect to be wrong in this prediction.

(Tom McLaughlin)

Serverless will become more normalized throughout the community. The big releases from AWS will be more focused on Serverless ML/AI services rather than data or messaging services. (John McKim)

We will start to see dramatic reductions in “cold start” times and the developer tooling around many FAAS providers will become much more solid. (David Wells)

Who wins the most from the adoption of serverless technology?

People who make the connection between technical effort not being an end goal, but as a means of producing value.(Tom McLaughlin)

“Consumers!” — Yan Cui

“Developers!” — Raymond Camden

“Everyone!” — Elliot Forbes

“Should I tell that reddit user that serverless actually uses servers? Lol, what? — of course I should!”

Developers! The close second winner would be companies getting the competitive edge by not spending too much time on DevOps… unless they’re super cool and choose to use Kubeless. (Adnan Rahic)

In my opinion, everyone wins — individual developers, small-medium businesses and large enterprises alike. (Rupak Ganguly)

The devs using it.

Having to not worry about scaling, load balancing, security patches etc. make serverless a fantastic way to build robust applications with minimal DevOps. (David Wells)

Who has the most to lose?

PaaS offerings. While I do not see them going away, I definitely see using those options less. Of course, companies that offer both will have their bases covered. (Raymond Camden)

People and organizations who cannot adequately connect technical effort with value will lose the most. If you’re a company you’ll be outpaced by the competitor that produces better value quicker. (Tom McLaughlin)

Be cute, but don’t be stubborn.

The ones who don’t want to embrace the change, and not giving a serious look at the benefits that serverless computing offers. (Rupak Ganguly)

The container world & tooling around containers certainly will start seeing the effects as serverless tech gains more prominence. (David Wells)

Vendors that deliver technology or services for elements lower down the stack like servers and containers. For example, Digital Ocean and Docker. (John McKim)

Will serverless computing become the norm? When?

In some context, it will! But not every problem you face as a software engineer will be solvable by a single tool. Dedicated servers will still have their concrete use cases. (Adnan Rahic)

Yes, serverless computing will become the norm within 7 years. (Yan Cui)

Yes! Why do I think it’ll become the norm? It’s watching the Kubernetes community develop serverless frameworks. It’s watching Red Hat focus resources on Apache OpenWhisk. The infrastructure and platform management communities aren’t stopping at containers.

The question as I saw it a year ago was, “Do you adopt containers or do you adopt serverless?” It was a question about what technology to adopt and competing technical views. Seeing the serverless progress in the infrastructure management space makes me realize the future question will be, “Do you build your cloud platform or uses someone else’s?” Do you build or buy?… Which is just the same old basic question we’ve always been asking ourselves. (Tom McLaughlin)

We have seen the cloud evolve since 2009, but serverless computing is how the cloud should have been in the first place. (Rupak Ganguly)

Yes. But, this will take a long time. Many companies are still unsure about the cloud or serverless due to a lack of understanding. Advocacy and education will play key roles in driving Serverless to become the norm. (John McKim)

I think there’s a few ways to look at this question. One is — when will it become less the “hot new thing” and “just another option”. I think that’s becoming reality now. Serverless still has an incredible amount of hype around it, but in general it feels like it is becoming mainstream and less bleeding edge. That’s a great thing.

The second way of looking at the question is in regards to serverless becoming the default for application development. I think serverless has some incredible benefits over traditional server based architectures. While I don’t think serverless makes sense everywhere, I definitely see it as the option developers should check *first*. (Raymond Camden)

Serverless computing is going to revolutionize the way we build and deploy applications. (Elliot Forbes)

Yes. Serverless computing will become the de-facto way folks will build apps in the future. At Serverless Inc, we are already seeing large enterprise companies running massive workloads through tiny lambda functions. (David Wells)

How to get more people involved and interested in serverless methodology?

Proper education! Get the major players in the field to work together towards a common goal through organizing workshops, courses, and events.

Here’s an idea, just as DigitalOcean and GitHub have been collaborating to create Hacktoberfest as a way to get the community to contribute to Open Source, why not do something similar with the goal to educate people about serverless!? (Adnan Rahic)

More people sharing real-world experiences (both gains and pains) of their own adoption. (Yan Cui)

We need to talk and write more about it. Businesses and developers alike should share real use cases and publish case studies for others to learn from.

The influencers and early adopters need to share code, examples, advice, learnings and their experiences.

We need to unite as a community beyond our own products, irrespective of cloud platforms, and evangelize the business values & benefits of serverless architectures to the masses. (Rupak Ganguly)

Everyone agrees — we need to unite and educate!

I think the easiest way to get people started with serverless is to have them identify something non mission critical to port into a FAAS provider of their choice. Once they see how easy it is, they can often take that POC and expand it into many facets of their organization.

Approaching upper management with a proposal for a big bang rewrite on any tech is typically shut down within seconds. It’s much easier to start small, see the benefits from a smaller project, and expand serverless tech into other areas of the business. (David Wells)

The companies that are already on the cloud and run existing workloads on cloud providers like AWS, Azure and Google cloud have an upper hand. They are already past the “Why should I move to the cloud?” hump, and more aligned to take the next leap to going serverless.

It is not too late for the ones that are still not on the cloud — they should adopt the “Serverless First” mantra, and evaluate incorporating serverless technologies starting at the periphery of their systems. (Rupak Ganguly)

The best serverless demos make use of cats. Anything else is simply wrong. (Tom McLauglin)

Conclusion and gratitude.

I want to thank all the people who took the time to answer me and share their ideas (and also the ones who couldn’t take part for various reasons but gave their well-wishes and were very interested in seeing it published).

As mentioned in the beginning of the article, this is only a summary of the interviews, definitely go check out the full versions in Dashbird’s blog!

NB: You can follow me in Twitter and don’t forget to check out Dashbird.io for serverless monitoring and debugging!