Torque Cloud Framework is the foundation for the next stages of cloud software development. Unlike the current explosion of new tools and services, Torque Framework does not add to the complexity of cloud-native development and deployment. Rather, it transforms the complexity of cloud-native systems, from a liability into a new capability; from a cost to be mitigated and managed, to a resource that can be mined and fashioned into entirely new ways of developing software.
In this vision statement, we’ll outline each of the next five stages of cloud software development as we see them, and how new capabilities made possible by Torque Framework will facilitate their arrival. Our hope is that this exercise proves useful both for understanding the implications of our technology, but also as an invitation to co-create the future with us. Subsequent articles will treat each stage in-depth, with more technical context.
Now, back to the audacious statement at the top of this page. We did not arrive at this idea lightly. Claiming it will be possible for applications to self-generate is more than a fantastical brainstorm. Today, cloud systems are managed at the infrastructure level. Torque Framework reduces complexity by handling the minutiae of infrastructure, so developers can focus on architecture design. Dozens of platform-as-a-service solutions, like Heroku and Zeet, already accomplish a similar abstraction, but only in limited capacity, through closed platforms.
Once infrastructure becomes a natural reflection of architecture design, the industry will demand more from application functionality. Again, there are currently dozens of low-code and no-code platforms that attempt to shortcut this abstraction, but they serve only limited use-cases.
In contrast, our open-source framework, powered by libraries, provides a common foundation upon which any new products, services and business models can be built, by anyone. Because it’s an open framework, and not a closed platform, new technologies can be incorporated as new libraries. We see a virtuous circle emerging that will enable the global community of developers to spin out their own new products in ways we can only imagine today (and likely more products we struggle even to imagine).
Furthermore, it follows that a single framework, which can BOTH represent system components as libraries AND deploy them, stands to serve a critical role in actualizing how artificial intelligence will graduate from simple tasks, like suggesting lines of code, to assembling and deploying highly functional systems.
The next stages of cloud software development as we see them are:
- Self-Configuring Cloud Systems: architecture design fully de-couples from infrastructure
- Composable Subsystems: functional parts grow from atomic to systemic
- Functional Capabilities Beyond Infrastructure: libraries also deliver application functionality
- Self-Composing Architectures: application functionality fully de-couples from architecture design
- Self-Generating Applications: Enabled by AI, users generate their own applications
Let’s add a bit more color to each one for clarity and to see how one enables the next.
The first stage is already underway with the release of Torque Framework. Today, architecture is still coupled to infrastructure in limiting ways. Completely de-coupling design from infrastructure, the way Torque does, allows workloads to roam freely, with infrastructure changes a natural, automatic by-product of design changes. This capability raises developer work to a higher level, and the notion of configurations will fade into the background, the same way hardware and even virtual machines have.
As infrastructure becomes more and more de-coupled, stages two and three will take shape. We will move from designing and managing atomized services into the age of composability with modularized subsystems and application functionalities. Think of it as a "build your own web application" out of pre-existing building blocks, without the restrictions of a closed platform, or the limits of web services. Torque Framework was built to be composable, with this future state in mind.
Stage four brings the abstraction of the architecture layer, raising focus to the application level. At this stage, we will have fully modularized application development solutions. Custom code will be minimal and serve only to glue application modules together. Again, details of the lower level– this time architecture– will command less attention.
The last of these five stages will emerge over time, as a continuation of AI-assisted software development that will evolve concurrently with the rest. The reason we put AI-enabled applications last is a matter of stepping through each stage of complexity, packaging lower level details into ever-greater capabilities. We envision AI assuming the capabilities created in each previous stage, until it can operate at the highest level of generating applications.
While Torque Framework is already capable of bringing the first three stages to fruition, broader industry adoption will be necessary to make the complete vision real. It will likely happen in fits and starts; progress is never linear. But it won’t take long. Perhaps a decade or so.
As an industry, are we ready? Cloud-native software development, while nearly ubiquitous today, is still in a formative stage. Far too much expertise and manpower is required to wrangle the cloud beasts we build. We obsess about the cutting edge in operations, because it’s so hard. There is no doubt cloud-native technologies have introduced us to a new transformative phase in human-computer interaction that we have yet to fully realize.
“Human-computer interaction” as a term is instructive here, because it helps us expand our frame of reference to make room for what’s coming. This is not just about software teams gaining a new capability. It’s about rethinking our relationship to technology. What it means both to be a developer and a user will change when we finally unleash the cloud’s full potential to self-generate new applications, on demand, enabled by AI.
In upcoming articles, we will delve deeper into each stage of cloud software development, what becomes possible, and what we think it means for the industry (links will be added as new ones are published):
- The Current State of Cloud-Native Software Development (Torque’s Vision Part 2 of 7)
- Stage One: Self-Configuring Cloud Systems (Torque’s Vision Part 3 of 7)
- Stage Two: Composable Subsystems (Torque’s Vision Part 4 of 7)
- Stage Three: Functional Capabilities Beyond Infrastructure (Torque’s Vision Part 5 of 7)
- Stage Four: Self-Composing Architectures (Torque’s Vision Part 6 of 7)
- Stage Five: Self-Generating Applications (Torque’s Vision Part 7 of 7)