Premium
SU‐GG‐T‐394: Efficient Mass Deployment of ROIS Client Computers
Author(s) -
Kim DYJ
Publication year - 2008
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.2962144
Subject(s) - upgrade , software deployment , computer science , client , operating system , domain (mathematical analysis) , software , hospital information system , world wide web , software engineering , server , information system , engineering , mathematical analysis , mathematics , electrical engineering
Purpose: To develop an efficient method of deploying 31 new client PCs obtained as part of a Radiation Oncology Information System (ROIS) upgrade. The major challenges were 1) reconciling the Windows domain environment required by the ROIS, with the Novell environment required for institutional (hospital‐wide) applications, 2) getting all departmental, institutional, and extra‐institutional client applications to work properly under normal user privileges, and 3) the short weekend timeframe of the upgrade. Method and Materials: The 31 Dell GX755 clients were preinstalled with a standard setup from the hospital IT department, including Windows XP, Novell Client, and hospital‐wide applications. Deployment of the clients during the ROIS upgrade occurred in 2 groups. In Group 1, Windows domain membership and ROIS client applications were manually installed at each client. For Group 2, a single model client received the Group 1 installation, followed by exhaustive client testing. After fixing the various non‐ROIS client applications that were discovered to be incompatible with the ROIS setup, the “perfected” model client was cloned via Microsoft Sysprep and Knoppix Linux to the other Group 2 clients. Results: Group 1 clients were deployed during the ROIS upgrade weekend; however, nonfunctional hospital applications were subsequently discovered, requiring individual patching. Group 2 clients were deployed one week after the upgrade weekend, but all client software worked properly from the start. Conclusion: The Group 2 method of “perfect then clone” is the preferable solution for mass deployments as it automates numerous tasks that would otherwise require repeating on every client. Post‐upgrade, this method is undoubtedly superior as it provides an optimal method for restoring a disabled client. However, it requires that all troubleshooting occur “up front” before cloning, which, with proper planning, support, and expertise, should be possible to accomplish within the time‐pressured ROIS upgrade weekend.