An alternative technology prediction list for 2021

01 intro prediction

As Bing Crosby and Perry Como would say: now is that special time of year when we can take stock of where the software industry has witnessed major developments, stand back and understand the weight of platform level shifts… and consider major use case changes that could provide us with some insight into how the various classes of software application development engineers need to think about IT going forwards.

Actually, Bing Crosby and Perry Como would more likely say: it's beginning to look a lot like Christmas. But both software prediction stories and 1950s 'crooners' seem to surface at this special time of year, so let's join in the holiday spirit in a year that almost every living soul on the planet will be happy to see disappear.

So what can we expect for software in 2021?

The 'obvious' answer for key trending tech would be more cloud-native technologies with a heavy emphasis on hybrid clouds, all driven by the advantages of AI and process automation for autonomous control. But that's a fairly ubiquitous statement, so can we look for deeper granularity and specific next steps? If we had to make a slightly more alternative software predictions list for 2021, then what would be on it?

Trend #1: Cloud computing gets the disposability religion

With so many parts of our enterprise IT stacks moving to cloud and all its abstracted advantages, the push for observability will continue to be key. This is good news for the Application Performance Management (APM) players, the log file analytics specialists and the Continuous Intelligence platform companies.

But although cloud is not old or legacy software (yet), we can still say that some cloud instances are no longer suited to the changing needs of the use case they were spun up for. Some clouds will be growing old enough to be ranked as increasingly redundant and badly positioned to integrate with current IT department requirements. With cloud-based microservices further fuelling the creation of IT fabrics that perform specific niche functions, the need to kill off and retire code that is generally no longer fit for purpose becomes more pressing.

While disposability is a dirty word when we're talking about polystyrene coffee cups and plastic bags, cloud disposability is an environmental positive as it allows us to turn off (or, more likely repurpose) machines in our datacenters that can be tasked with new jobs. The responsibility today is for systems architects and cloud engineers to make sure that routes to cloud disposability are accessible and easy to execute.

Trend #2: Containers get ruggedized

Also featuring in the usual suspects category for key trending tech are containers. A means of encapsulating all the application logic, operating system elements and supporting framework technologies into one discretely defined package, containers are a key part of cloud growth along with Kubernetes as an orchestration platform to direct them. Tim Mackey, principal security strategist at the Synopsys CyRC (Cybersecurity Research Centre) notes that container images can be used 'as is' to provide a shared service, such as with a database, or will be embedded as a base image to create a new container image.

Going forwards, Mackey advises that we will start to focus more carefully on the provenance of containers. This is important because the design and security practices of the team creating the original container image have a direct impact on the security of the resultant system.

"Implicit [container] trust is risky from a security perspective, which is why many organizations are now creating hardened container images where the image hardening process is managed by a dedicated team skilled in operating system 'ruggedization', which is separate from the core development team," said Mackey. He notes that these hardened images are then pushed to an internal registry and policies are defined that only allow images originating from hardened images in that internal registry to execute in a production cluster.

Trend #3: The microservices bubble bursts

Microservices have been all the rage. Enterprise IT shops are busily deconstructing and rearchitecting applications for microservices so that they can build more composable technology stacks. But could the popularization bubble be about to burst? After all, the more granular we go, the more operational data management challenges we face. Could we face the stark reality that, actually, microservices end up being just as weighty a beat to take on as monoliths?

But how do we end up like this in the first place? A key driver is thought to be the fact that technologists often want to use the newest and most esoteric tools, but these are often the most untested and difficult to implement correctly.

Founder of software engineering company Indorse Gaurang Torvekar, says that the problem is, there is just as much chance of getting into a tangle with microservices as there is with monoliths. "That's because people often start breaking things up into microservices too early - they've heard it will be a good idea, they've seen some of the benefits, so they try to apply it to everything. In doing so, they create what some might call a micromess, a micromuddle or perhaps even a micromonolith," said Torvekar.

Trend #4: AI will go one level deeper

We know that Artificial Intelligence (AI) has finally come of age and that Machine Learning (ML) is part of the new fabric of enterprise IT. We also know that AI will help drive low-code no-code platforms to enable us to build applications more quickly. The problem is, you can't just stick an AI autopilot on a broomstick and expect it to know how to clean the house. We have failed to think about the lower level 'feature engineering' where large, complex and sophisticated complex AI/ML workflows really work at the heart of data science.

Founder & CEO of dotData Ryohei Fujimaki says that the vast majority of the work that data scientists must perform is often associated with the tasks that precede the selection and optimization of ML models.

"This means that organizations will need to look for new, more sophisticated automated machine learning platforms that enable true automation, from automatically creating and evaluating thousands of features (AI-based feature engineering) to the operationalization of ML and AI models -- and all the steps in between," said Fujimaki. 


Trend #5: Hyperautomation will go hybrid

We've built autonomous software controls, okay, we get it. We can now perform AI-driven Customer Relationship Management (CRM) recommendations to offer us pizza discounts on Tuesdays; we can auto-set databases to patch, update and run system maintenance tasks while we slip off down to the sports bar; and we can use Robotic Process Automation (RPA) bots to shoulder all the donkey/grunt work associated with an increasing number of enterprise tasks across a variety applications. Yes, we do know that part.

What we're perhaps not thinking about—and what may deliver an epiphany moment in 2021—is the need to push all that automation to be able to work as a more expansive source of multi-platform multi-environment autonomy. UiPath executive vice president of product & engineering Ted Kummert says that we're on the point of platforms now infusing AI to deliver hybrid hyperautomation.

"This is a combination of solutions, from process mining to document understanding, that will enable organizations to expand use cases for automation across the business and to more complex workflows. It will also be increasingly critical that automation platforms can be deployed in hybrid models – meaning SaaS, IPaaS and/or on-premises. To do so, vendors will need to deliver consistent APIs, Ui experiences and reference architectures across delivery modes," said Kummert.

Trend #6: Cloud kitchens will cook up the edge

As we have already discussed here, computational storage is already confirmed to be one of the tech darlings of 2021. Being able to perform compute functions at the storage plane and not having to transport data back and forth to the CPU can have a profound impact on the way we build the Internet of Things (IoT) at the 'edge' going forwards.

Computational storage and the increased use of Solid State Disk (SSD) technologies may never be considered to be at the raciest end of the IT spectrum. After all, people would still rather talk about user interfaces, new mobile devices and all manner of AI. But when we see new edge-based computing power with AI and Machine Learning (ML) intelligence, then we can start to allow machines to really 'see' the world and develop new levels of computer vision understanding.

"Thinking about computing at the edge, mobile-powered e-commerce will continue to make giant strides in 2021," said Dheeraj Pandey, founder, CEO and chairman of Nutanix. Pandey says that we will now come to know the 'remote consumer' far more prevalently – and this will drive technology to develop for online product distribution centers and so-called 'cloud kitchens' i.e. physical kitchens that work for delivery-only, with no restaurant (and often no windows).

So it's going to be a question of remote computing empowerment in so many ways, but says Pandey, this will drive the need for massive back end reliability, availability, security, elasticity and performance requirements in server rooms and ROBO (remote office, branch office) computing closets.

"In light of these developments, the cloud will further disaggregate and miniaturize at rapid speeds in the coming year. Containers and data pipelines will become ubiquitous at the edge, with PaaS platforms available in even single-server clouds. Also, governance around images and video data being produced at the edge will bring even more meaningful applications of AI and ML in the hybrid enterprise," said Pandey.

Trend #7: Post Covid-19 fallout & our quantum future

Of course we've only just scratched the surface here as we have attempted to follow a few specific trending discussion threads and think about what they are going to mean for enterprise IT systems development. We've been from microservices to hyperautomaton and cloud orchestration without even mentioning blockchain, or more accurately, blockchains plural (both public and private) and their obvious growth factor into 2021 and beyond.

In the year 'after' Covid-19, the tech industry and the world of work at large will clearly continue analyzing the use of remote working environments. There are no prizes for talking about low-code Rapid Application Development (RAD) in the wake of pandemic-driven corporate realignments, the t-shirts have already been printed for that trend.

Looking to 2021 and predicting our key software application and data services shifts is easy in one sense i.e. we know there's a lot of catch-up, recovery and re-architecting to do. Looking to 2022 and perhaps a renewed focus on the future promise of quantum computing is much tougher.

We're just about comfortable with talking about petabytes and then exabytes (given that you probably have a terabyte or two in your pocket by now), so when the zettabyte discussion starts happening, then you'll know we're really looking to the future. Until then, keep washing your hands.