Storage technologies are entering the market with performance vastly superior to conventional storage devices. This technology shift requires a complete rethinking of the software storage stack.

In this talk I will give two examples of our work with Optane-based solid-state (block) devices that illustrate the need for and the benefit of a wholesale redesign.

First, I will describe the Kvell key-value (KV) store. The key observation underlying Kvell is that conventional KV software on fast devices is bottlenecked by the CPU rather than by the device. Kvell therefore focuses on minimizing CPU intervention.

Second, I will describe the Kvell+ OLTP/OLAP system built on top of Kvell. The key underlying observation here is that these storage devices have become so fast that the conventional implementation of snapshot isolation – maintaining multiple versions – quickly leads to the device filling up. Kvell therefore focuses processes new versions as they are created.

This talk describes joint work with Oana Balmau (McGill University), Karan Gupta (Nutanix) and Baptiste Lepers (University of Sydney).

---

Please contact the MPI-SWS Office Team for the ZOOM link information. .

Machine learning (ML) based algorithms assist human decision-making in a variety of scenarios, ranging from medical diagnostics to bail decision-making. The potential societal impact of using machine decision aids in real-world settings sparked concerns about their accuracy and fairness, and inspired a flurry of research on algorithmic fairness, accountability and transparency. However, in many settings, algorithms do not make decisions, but only assist human decision-makers. In this thesis, we go beyond studying the fairness and accuracy of decision aids, and study machine-assisted human decision-making as a whole. Specifically, we study how people perceive and utilize machine decision aids.

Please contact the MPI-SWS Office Team for the ZOOM link information.