Search

Toward machine learning based prediction services, the prediction server has multiple predictors and selects an appropriate one based on past feedbacks from the clients. In this case, three messages including request, reply, and feedback, are required for each prediction request. Packets are typically transmitted and received via a network protocol stack in OS kernel, and performance improvement can be expected by avoiding the protocol stack since it degrades the communication performance especially for small packets. We implement the prediction server using network optimization approaches including kernel-bypassing and in-NIC processing approaches. Evaluation results show that these network optimizations are beneficial to improve the prediction server performance compared to a baseline prediction server using a standard network protocol stack.

To realize technological innovation based on advanced university technologies, it is important to establish collaborative research projects between universities and industry in the early stages of R&D. One way to stimulate the creation of university-industry collaborative research is to have a systematical coordinating mechanism between a university and industry. We analyzed the effectiveness of a university-industry coordination program at creating university-industry collaborative research projects. We used data from the matching program called Proprius21 at the University of Tokyo and found out that the matching program has increased collaborative studies at the University of Tokyo. We also analyzed the R&D characteristics of various firms using the relation between the firm's R&D intensity and the range of firm's R&D areas. The matching program is especially effective for large companies in ICT-related fields, although it is not so effective for efforts involving life science technologies, which are overall the most active area for university-industry collaboration. Results show that the matching program is most effective for creating collaborative studies with technology-oriented firms engaged in various business domains.