Andre Pereira1, Mostafa Atri2, Patrik Rogalla3, Thien Huynh3, Martin E O'Malley4. 1. Department of Medical Imaging, University of Toronto, 17-525 Balliol St, Toronto, ON, M4S 1E1, Canada. andrepereira71@gmail.com. 2. Department of Medical Imaging, University of Toronto, 585 University Avenue, Csb_1C-569-TGH, Toronto, ON, M5G 2N2, Canada. 3. Department of Medical Imaging, University of Toronto, 585 University Avenue, Ncsb_1C-571-TGH, Toronto, ON, M5G 2N2, Canada. 4. Department of Medical Imaging, University of Toronto, 610 University Avenue, Room 3-920 - PMH, Toronto, ON, M5G 2M9, Canada.
Abstract
PURPOSE: The value of a teaching case repository in radiology training programs is immense. The allocation of resources for putting one together is a complex issue, given the factors that have to be coordinated: hardware, software, infrastructure, administration, and ethics. Costs may be significant and cost-effective solutions are desirable. METHODS: We chose Medical Imaging Resource Center (MIRC) to build our teaching file. It is offered by RSNA for free. For the hardware, we chose the Raspberry Pi, developed by the Raspberry Foundation: a small control board developed as a low cost computer for schools also used in alternative projects such as robotics and environmental data collection. Its performance and reliability as a file server were unknown to us. For the operational system, we chose Raspbian, a variant of Debian Linux, along with Apache (web server), MySql (database server) and PHP, which enhance the functionality of the server. A USB hub and an external hard drive completed the setup. Installation of software was smooth. RESULTS: The Raspberry Pi was able to handle very well the task of hosting the teaching file repository for our division. Uptime was logged at 100 %, and loading times were similar to other MIRC sites available online. We setup two servers (one for backup), each costing just below $200.00 including external storage and USB hub. CONCLUSION: It is feasible to run RSNA's MIRC off a low-cost control board (Raspberry Pi). Performance and reliability are comparable to full-size servers for the intended purpose of hosting a teaching file within an intranet environment.
PURPOSE: The value of a teaching case repository in radiology training programs is immense. The allocation of resources for putting one together is a complex issue, given the factors that have to be coordinated: hardware, software, infrastructure, administration, and ethics. Costs may be significant and cost-effective solutions are desirable. METHODS: We chose Medical Imaging Resource Center (MIRC) to build our teaching file. It is offered by RSNA for free. For the hardware, we chose the Raspberry Pi, developed by the Raspberry Foundation: a small control board developed as a low cost computer for schools also used in alternative projects such as robotics and environmental data collection. Its performance and reliability as a file server were unknown to us. For the operational system, we chose Raspbian, a variant of Debian Linux, along with Apache (web server), MySql (database server) and PHP, which enhance the functionality of the server. A USB hub and an external hard drive completed the setup. Installation of software was smooth. RESULTS: The Raspberry Pi was able to handle very well the task of hosting the teaching file repository for our division. Uptime was logged at 100 %, and loading times were similar to other MIRC sites available online. We setup two servers (one for backup), each costing just below $200.00 including external storage and USB hub. CONCLUSION: It is feasible to run RSNA's MIRC off a low-cost control board (Raspberry Pi). Performance and reliability are comparable to full-size servers for the intended purpose of hosting a teaching file within an intranet environment.
Authors: Steffen Möller; Hajo Nils Krabbenhöft; Andreas Tille; David Paleino; Alan Williams; Katy Wolstencroft; Carole Goble; Richard Holland; Dominique Belhachemi; Charles Plessy Journal: BMC Bioinformatics Date: 2010-12-21 Impact factor: 3.169
Authors: Daniel Barker; David Ek Ferrier; Peter Wh Holland; John Bo Mitchell; Heleen Plaisier; Michael G Ritchie; Steven D Smart Journal: BMC Bioinformatics Date: 2013-08-12 Impact factor: 3.169