Sector: A high performance wide area community data storage and sharing system

We introduce Sector, a system that provides data storage and sharing services over high performance wide area networks. There is now a critical need for such systems, given the rapidly increasing sizes of scientific datasets. Meanwhile, developing such systems is becoming practical because of the rapid growth of optical networks. The goal of Sector is to serve a community where users can upload and store large datasets, easily share their data with others, and perform distributed data processing over it using a very simple API.Sector uses a peer-to-peer routing mechanism to organize the participating nodes. Sector servers provide very simple and basic functions for storing, locating, and accessing data. Users can use a Sector client to access data or use the Sector client API to write distributed applications. We have successfully used Sector to store and distribute various products from Sloan Digital Sky Survey (SDSS) to astronomers around the world.