Scalable Abstractions for the Paradigm-Agnostic Web

Scalability of today’s Web technologies is limited by their dependence on the client-server paradigm: content and services reside on servers in data centers, and the end users’ client computers do not directly interact with one-another. I would like to enable a paradigm-agnostic Web, in which the clients could play an active role as an extension of the cloud, and where Web applications could take advantage of the full spectrum of distributed computing mechanisms, e.g., peer-to-peer overlays, gossip, and distributed replication. In paradigm-agnostic Web applications, the underlying communication and storage architecture could be selected automatically by a compiler or at runtime, to fit the target environment or workload patterns.

In this talk, I will introduce Live Distributed Objects (LDO), a new programming methodology and a platform designed for the paradigm-agnostic Web. I will also briefly talk about three closely related technologies: a formalism and a programming language for scalable distributed protocols, a scalable multicast substrate, and a concurrency abstraction for multicore CPUs.

Speaker Details

Krzysztof Ostrowski is a Postdoctoral Associate at the Department of Computer Science at Cornell University. He received his Ph.D. in Computer Science from Cornell in 2008. Prior to joining Cornell, he spent a few years in the industry. His research is focused on programming languages and distributed architectures that enable scalable Web applications. He created and is actively developing the Live Distributed Objects programming model and open-source platform. He published and served as a PC and journal reviewer in Web, distributed systems, and programming language communities.

Date:
Speakers:
Krzysztof Ostrwoski
Affiliation:
Cornell University
    • Portrait of Jeff Running

      Jeff Running