Google File System

Behind apparent simplicity of Google applications, there is a plethora of technologies. Developed by many engineers over many years, these technologies allow Google to achieve ambitious goals quickly and efficiently. Today we will discuss one of the major Google technologies, called Google File System or GFS. Google applications, from search to gmail, work in close to real time and at a massive scale. Therefore, Google applications need a scalable, massively distributed, and fault-tolerant file system that allows space and time efficient data storage and retrieval. In that regard, GFS shares the goals of other distributed file systems. However, GFS design is driven by the key observations about Google work load, and thus, can provide specific support to Google applications. Over years, GFS has successfully met storage needs for Google. Large GFS clusters can store terabytes of data, and are concurrently accessible by hundreds of clients. We will talk about some performance benchmarks at the end of the talk. In addition, we will discuss career opportunities for young engineers at Google. For possible job opportunities, you are advised to bring your resume and chat with Dr. Ray after the talk.

SPEAKER: Sibarata Ray, received his Ph.D. from University of Nebraska-Lincoln in 1992, majoring on Computer Science. He is currently working at Google Labs in Los Angeles, California.

Place: Fenton Hall, Rm. 105.

Date and Time: OCtober 24, 4 pm

Sponsored by: MACS Scholarship Program