Energy-Efficiency Scheduling For Delay- Sensitive Stream Mining Systems

Mentor(s):

Dr. Shaolei Ren

Instructor:

Masoud Sadjadi

Description:


Energy efficiency for data-stream mining systems is a major concern. In a weakened economy, major companies struggle in an effort to provide the same quality while saving costs. This, in of itself, is a tremendous challenge. How does an entity lower necessary costs while having no alternate way to scale down its operation? Due to various functional requirements of the system in whole, scaling down is nary a possibility. Taking an example of a large online business ( say, Amazon ), powering their operations requires several high-energy, high-strength servers to maintain a reasonable time constraint. There would be no true point to use an online retailer if it took the same or double the time to make a purchase. Likewise, a video-streaming site like YouTube would not be so desired if the entire streaming process reminded people of accessing their internet content through a 56K modem. Needless to say, the gravity of the situation rears its ugly head. Without having to sacrifice quality, a solution could emerge within the infrastructure of how work is done. This introduces the problem and inception of a project focusing on energy-efficient scheduling for delay-sensitive stream mining systems. As aforementioned, traditional hardware setups are configured as a colony of powerful servers working in tandem to offer the quickest service possible, usually at the expense of the bank accounts of said companies. Granted, it could be said that the average income greatly nullifies this as a concern, but as data grows in size and complexity and consumer demand for faster speed increases nearly at an exponential rate, the concern to meet these same expectations within reason grows as well.

Team Members

Frank Fernandez

John Rodriguez