2020: A Developer’s Preview
Interesting Trends in Devtech
The worldwide public cloud services market was projected to grow 17.5 percent in 2019 to total $214.3 billion, up from $182.4 billion in 2018, according to Gartner, Inc.
Although $214.3 B is quite a large industry, the growth factor is the eye-catching feature.
Cloud growth is projected to continue on its expansion track, arriving at almost $300 Billion USD revenue during 2020 (Forrester).
That just means one huge thing to you and me: opportunity.
But What Exactly is It?
There are really two kinds of cloud software: end-user and developer. SaaS and IaaS.
The *aaS acronyms start to just be a muddle. For you and I who know, there is devtech, and then there is everything else.
The question is are you building software to deliver services to end-users, or are you building tools that developers use for that purpose?
In either case, the market is expanding fast.
The Garner analysts identify four chief areas in cloud computing:
- IaaS: Virtualized infrastructure
- PaaS: Runtime abstractions like Cloud Foundry
- SaaS: The end-product of most development: a service consumed over the ‘net
- BPaaS: A catch-all descriptor for higher-order SaaS designed to encapsulate entire business processes
That is to say, basically: two kinds of devtech (IaaS and PaaS), and then all the end user SaaS stuff.
In truth, there is an enormous range of devtech that I’ll call devtools. Not really I/PaaS, but still devoted to development. (Think github.)
Anyway, however you slice it, there are some major trends afoot affecting how these industries are evolving. (I am JACKED UP about the potentials.)
A Programmer's Guide to Creating an Eclectic Bookshelf | Data Driven Investor
Every developer should have a bookshelf. The possible set of texts in his cabinet are myriad, but not every collection…
Here’s my take on what is going on and how it’ll play out in 2020.
AI Cools Off
Did you know artificial intelligence has gone through several boom and bust cycles?
They call the cool-down periods “AI winters”.
We are looking at that now.
Here’s what happens:
- New techniques and expertise build up, and more advanced uses for machine learning arrive on the scene.
- Word spreads and the marketing folks pick it up.
- The hype begins and people become expectant of true, Earth-shaking computer-based intelligence
- The hype goes off a cliff of unreality as the actual work of machine learning pushes incrementally forward
There are obviously valid and powerful use cases for machine learning. They are however solidly in the realm of what we’d normally think of as software-soluble problems anyway.
Yes, it is interesting that software can help me start my customer service interaction. No, it isn’t outside that range of what I’d expect software to tackle.
In short, machine learning is another set of useful techniques for attacking software problems.
What it is not is about to spawn human-rivaling creativity.
Not even close. Moreover, it is not ever going to usher in artificial consciousness, an idea that lingers in the backdrop of the whole discussion.
We might make software that passes the Turing test, ok. But that test itself really passes the buck on the deeper questions around AI.
The long and the short of this discussion, as we skirt around the philosophical nature of it, is this: the hype around AI in the industry is beginning to fade.
Looks like human beings will have to continue doing the real thinking.
Yes. I’m aware that I will be lambasted for this one. Mostly by bots.
IaaS is Homogonized, New Battlefields Sought
IaaS grew to $32.4 billion in 2018, and 2019 is on course to continue that trend. Web apps were already a rapidly expanding concern, and cloud tech just threw a bunch of gasoline on it.
IaaS platforms like AWS’s EC2 have gone from a revolution, to a critical part of any big tech companies offerings, to a homogenized field of offerings.
That trend will continue in 2020. That sets the stage for ongoing developments in the industry, because that is the baseline competition in cloud development: who provides the hosting and tools for developers.
All this means the horse race is shaking out, the market is fairly stabilized. Competition is beginning to focus on pricing. (Google Cloud Platform just announces a 33% price cut in Dec. 2019). Offerings like Vultr are becoming attractive.
The field is not marked by vendor lock-in as feared so much as it is marked by vendor indistinctness. Little differentiates the competitors.
Much of the IaaS space, like spinning up a VM instance and routing soft-network to it, building serverless functions, deploying monitoring solutions, running PaaS apps, datastore hosting, kubernates and similar offerings — much of this looks awfully similar between the major cloud providers.
Indeed, there is an explosion of offerings and ways of relating not only cloud services but dealing with the interaction of coding and its internals and processes with IaaS.
The net result is NOT a simplification of dev or devOps. In fact, its a compounding ramification of complexity.
A developer is faced with a complex array of options and questions. Not to mention just answering simple questions like: “What am I being charged for?”
New Dev Tools
The big question is where is the next battleground?
The answer is: in devtools that can digest and make manageable the wide array of offerings; digest the rapid introduction of new techniques and tools; and make it higher-order controllable by developers.
I’ll squish this all under the heading of “new devtools.”
Watch for it. It’s coming and will alter the face of development.
A key current that will underlie coming advancements is…
That is: we don’t really care which cloud provider we are using: we just need disk, ram and network. Now run it.
How far can cross-cloud-platform approaches be pushed?
Just how abstract can we make infrastructure? If one cloud provider cuts prices, what kinds of workloads can be moved efficiently off another provider?
How much can that be automated?
If we look without attachment at the cloud as a marketplace for services governed by price, quality and so forth, we get one picture, but when you consider that running software or persisted data have an economic inertia, you get another.
Even with advanced tools that can look at apps and infrastructure vendor-agnostically as simply capability providers, there is a limit to them.
But that being said, there is certainly the opening of new realms for dealing with infrastructure that bestrides cloud providers as simply what they are: competing services.
This one circles back around to asking what will cloud-providers do to distinguish themselves. The fact that a customer is already using a platform is nice, but the more that tooling improves to abstract the cloud providers offerings, the more these customers are vulnerable to poaching.
React Remains Dominant Front-End Framework
I had a rude wakeup where I realized that my underdog favorite, VueJs, wasn’t the best framework out there. ReactJs is.
That feelings has only been reinforced over time, and I wonder out loud if the sheer weight of Google backing Angular is the only thing that will make it long term competitive with React.
In short, React dominates the front-end landscape, and will continue to do so, until something new comes along usurp it’s place. It ain’t coming from the frameworks already in the race.
Something like lit-html, perhaps.
In any case, React remains the champ for 2020 until that next upstart comes along.
The one thing I can guarantee is that something will come along and obviate React.
It’s the one thing you can count on in software. Change.
We’ll see if the upstart arrives in 2020. I kinda doubt it, but you never know.
More Big-Corp Musical Chairs
It’s been interesting to watch the mighty tech giants scramble and acquire to survive the cloud revolution. It’s like musical chairs, each company concerned it won’t have a big enough footprint to survive.
It looks like the biggest players will still be around when the music finally stops, however, the jockeying will continue for some time.
This again resolves to opportunity for you and I. Why?
Because the deep pockets of the industry are still on the prowl for advantages obtained by buying up technology.
Virtualized resources are inherently more vulnerable, and micro-service architecture increases security complexity.
Compound that with the difficulty of managing cloud resource useage across large orgs, and there is the certainty of security remaining an on-going and critical area of development.
It doesn’t look like it. I don’t think 2020 will be the year that happens.
Look, you and I build software.
We like to build it.
We are in the midst of an absolute kaliedescopic hurricane of potential things to build. Things that can break new ground.
Things that you’ve thought of that if you don’t do it, it won’t happen.
Don’t ask where we are in the cloud revolution.
No, ask where we are in the software industry arc. Look at our current phase as more of the next big surge after the dot com breakthrough.
Fifty years from now, they’ll look at this wild ride and wonder what it was like.
It was bad ass.