Romulus:
As web applications expand over interconnected machines, fault tolerance becomes hard to guarantee. Highly available collections of systems require careful planning during software development, a costly process that is near impossible to apply retroactively. Romulus is a form of synchronization of multiple instances of Remus—a system described by Brendan Cully and his team for providing cheap and easy high availability of virtual machines (Cully, et al. 2008). Romulus seeks to provide easy and cheap high availability for clusters not seen in commercial products. Hardware limitations of today’s processors do not allow single machines to accomplish today’s workloads. As workloads spread over machines onto clusters, a system should also be available for providing reliability since multiple machines are more likely to fail than a single machine. With Romulus, high availability can be retroactively, easily, and cheaply applied to a cluster of networked machines.
Bio:
I am a student at The George Washington University working towards a BS in Computer Science with a minor in Communications. In addition to my coursework, I have also conducted research funded by the National Science Foundation and worked under the Summer Undergraduate Research Fellowship program at the National Institute of Standards and Technology (NIST) in Gaithersburg, MD.
Before relocating to Washington, DC, I spent my time in New Jersey. In my free moments, I enjoy ballet, running, and reading.
No Documentation