Skip to Main Content
We are entering an exciting era for systems design. We can expect the first exascale computer (with one million trillion operations per second) to appear around the end of the next decade. In addition to continued advances in performance, we are also seeing tremendous advances around improving power, sustainability, manageability, reliability, and scalability. Power management, in particular, is now a first class design consideration, and recently, system designs have gone beyond optimizing just operational energy consumption to examining the total lifecycle energy consumption of systems for better environmental sustainability. Similarly, the emergence of cloud computing, in addition to introducing a new exciting model for delivering computing, has enabled and demonstrated significant advances in scalability and innovations in the software stack. Looking further out, emerging technologies (such as photonics, non-volatile memory and 3D stacking) and new workloads (around data-centric workloads) offer interesting new opportunities. The confluence of all these trends motivates a rethinking of system design, one driven by these emerging future workloads and better leveraging hardware-software codesign for upcoming technologies. In this paper, we discuss these opportunities and examine what such a rethinking means for the basic systems’ building blocks of the future, specifically speculating on a likely design approach that we call “nanostores” and its implications.