Called BRASS (Building Resource Adaptive Software Systems), the ambitious project aims to create a software system that can remain robust and functional for over 100 years.
Whether at your office or at home, everyone knows what a pain it can be to upgrade software. Between lost time and the frustration of losing a desirable user experience, it’s no wonder that people are hesitant to make the switch to the latest and greatest software.
To further compound the pains of an impending upgrade, users have to live with the specter of increased vulnerability to cyber-attacks if an upgrade isn’t undertaken immediately. For government, military and security services this is doubly true.
Is there a way to build a computer system and software that doesn’t require upgrades? Even if there isn’t, DARPA is looking to find out.
While the notion of building a platform that doesn’t need an upgrade is probably as complicated as biological life itself, DARPA’s researchers believe they have an idea of how to design a system that evolves and lives in a digital ecosystem that’s as vibrant and anarchic as real life.
According to researchers, over the next four years the BRASS team will create a framework, and possibly a prototype, of a system that can adapt to changes in resources and its environment sans hardware and software upgrades while maintaining high-end productivity.
Of course, a radical departure like BRASS will require a completely “clean-slate approach” to software development. Engineers expect that new “linguistic abstractions, formal methods, and resource-aware program analyses” will need to be created from scratch.
Essentially, DARPA’s ambitious plan might be nothing short of a reinvention of computing as we know it. Not surprisingly, engineers at the advanced technology department grasp the gravity of the task before them.
“Ensuring applications continue to function correctly and efficiently in the face of a changing operational environment is a formidable challenge,” said Suresh Jagannathan, BRASS’s program manager. “Failure to respond to these changes can result in technically inferior and potentially vulnerable systems.”
Daunting doesn’t even begin to describe DARPA’s task. But what if they’re successful? What if BRASS leads to a real world machine?
Will it be as relevant to systems that aid the world of fundamental science or engineering as it could be for systems that demand a high level of security?
Without having to imagine too hard, I think the answer has to be an emphatic YES.
Follow me here for a second.
Now, if I’m reading the BRASS tea leaves correctly, if you were to slowly remove resources from a super-computing core, BRASS may continue to operate with equal efficiency, even as its resources become increasingly scarce. That means that anything from iterative design optimization to climate monitoring could be done relatively cheaply.
Going a bit further down the rabbit hole, you can imagine how an adaptive, upgrade-proof system is a giant leap toward artificial intelligence. If we follow the biology reference I made earlier in the article, it’s almost as though BRASS would be the “body” of an artificial intelligence before it gained the ability to make decisions for itself. BRASS is essentially the self-dividing, information copying, evolutionary template from which intelligence emerges.
While all of this may seem completely far-fetched (not to mention impossible), I think that it might be what DARPA’s team is working on.
Will BRASS have a major impact on engineering? If it’s a success it surely will.
Source: DARPA