top of page

Upcoming Events

We list upcoming events of interest to our members at the top of this page. If you are looking for recordings or resources from past events, you can jump to our past event listings below.

logo-05-academy-final-transparent-dark-bckg.png

Virtual Keynote on October 8, 2025 - 2:00 pm Eastern Time (UTC−05:00)

On Interplanetary and Relativistic Distributed Computing

Interplanetary distributed systems, such as the Interplanetary Internet, and the Global Positioning System (GPS) are subject to the effects of Einstein's theory of relativity. In this talk, we will formulate a unified computational model for relativistic and classical distributed computing systems and study the relationship between properties of distributed algorithms deployed on the two types of systems. Classical executions are totally ordered in time, whereas the steps of a relativistic execution are only partially ordered by the relation of relativistic causality. We relate these two physics-dependent execution types through a third—purely mathematical—notion of a computational execution, which partially orders steps by the relation of computational causality. We relate relativistic, classical, and computational executions of distributed algorithms through a central theorem, which states that the following are equivalent for any distributed algorithm ALG:    (1) ALG satisfies a property P classically;    (2) every relativistic execution of ALG satisfies P in every reference frame; and    (3) every total ordering of every computational execution of ALG satisfies P. As a direct consequence, we prove the equivalence of the standard, relativistic, and computational formulations of linearizability. Our results show that a host of algorithms originally designed for classical distributed systems will behave consistently when deployed in relativistic, interplanetary distributed systems. ​ Even as our results show that every classically linearizable distributed algorithm is relativistically linearizable, we demonstrate that different observers may necessarily have to disagree on the linearization order of operations. Our demonstration is through the derivation of a chronological invertibility inequality for a distributed system consisting of two satellites performing a pair of operations on a Jayanti-Tarjan union-find object whose computation is subject to observation by travelers flying in two spaceships. The inequality shows that with appropriate settings of the distance between the satellites, the speeds of the spaceships, and the computation times of the two operations, the two observers must see the operations happening in opposite orders. Finally, we instantiate the terms of the inequality to show that such chronological inversions are easily possible when the interplanetary system consists of satellites at the distance between the Earth and Mars and the travelers are flying at standard speeds of spaceflight.

600x800siddhartha.jpg

Siddhartha Jayanti

Siddhartha Jayanti is an Assistant Professor of computer science at Dartmouth College, where he leads the Distributed Computing and Verification Lab. His research spans algorithms, verification, and applications with a focus on designing simple, fast, scalable, and reliable solutions to challenging multidisciplinary problems. Siddhartha's work has touched several areas of inquiry, including distributed computing, algorithms and data structures, verification, relativistic physics, economics, security, and machine learning. Prior to Dartmouth, Siddhartha worked as a research scientist at Google Research, where he designed and deployed fast and formally verified algorithms for large-scale data processing and clustering. Siddhartha received his Ph.D. from MIT and bachelor's from Princeton. His MIT PhD dissertation on Simple, Fast, Scalable, and Reliable Multiprocessor Algorithms received the 2023 ACM-EATCS Principles of Distributed Computing Doctoral Dissertation Award.

bottom of page