A Library's Integrated Online Library System: Assessment and New Hardware Implementation
暂无分享,去创建一个
For more than a decade, a consortium of academic libraries in southern Nevada has shared a central integrated online library system (IOLS), Innovative Interfaces' Innopac (Innovative when referring to the vendor, Innopac when referring to the software). At present, this consortium includes the University of Nevada, Las Vegas (UNLV) (both the main university library system and the UNLV law library), the Community College of Southern Nevada, Nevada State College, and the Desert Research Institute. The last central-site server was purchased and installed in 1997. In the four intervening years, tremendous growth occurred with the system, necessitating a hardware upgrade. Prior to committing to a fiscally significant hardware upgrade, library management felt it prudent to conduct an analysis of the library-system-vendor marketplace in general, to validate the consortium's continued commitment to the existing vendor. This article discusses background assessment work and the subsequent planning and installation of the consortium's new central-site hardware. ********** The last central-site hardware for a consortium of academic libraries in southern Nevada was purchased and installed in the spring of 1997. At the time, the vendor indicated that the system purchased had enough memory and disk space to accommodate anticipated growth in users and data for the next three to five years. As reality sunk in that the libraries would need to upgrade the hardware, staff began looking at the true growth of the system, which proved to be quite substantial. The server's memory had already been upgraded to help accommodate the growth. Between 1997 and 2001, bibliographic, item, and patron record counts had increased by 59 percent, 86 percent, and 26 percent, respectively. The number of dedicated staff logins to the system (that is, potential simultaneous staff users) had more than tripled from thirty-three to 105. System log files associated with these accounts showed consistent use, at peak periods, of nearly all available logins. End-user statistics--circulation and interlibrary loan statistics, OPAC searches, and proxy-server use--varied, some rising and others falling. However, perhaps more telling than any statistic set, the libraries had added a tremendous amount of new software and routine responsibilities associated with the system, above and beyond regular vendor software updates. The University of Nevada, Las Vegas (UNLV) law library had joined the system, at which time additional acquisitions and serials units were installed. The library had installed an interlibrary loan module, an interface to the main library's automated storage and retrieval system (ASRS), self-check units, Z39.50 client-server software, and proxy-server software. Innopac, by Innovative Interfaces (Innovative), had introduced their new Java-based Millennium system, part of which had been installed. In addition to software, massive item record lists were regularly created for use in conjunction with the libraries' 3M Digital Library Assistant devices. The vendor's patron application program interface (API) module was installed and used for patron authentication purposes. Clearly, the consortium had stretched the boundaries in terms of what was expected from the central-site hardware and, for that matter, traditional integrated online library systems (IOLS) functionality in general. The effect of all this growth was obvious: lists and system backups took hours to create, staff menus took a long time to load--in short, things were slow. From a systems-analysis standpoint, could it be proved that the system was overtaxed and in need of an upgrade? Having obtained root-level access permission to the system, the library looked to measure system-load averages to provide some data-driven proof that the existing system was overburdened. An automated log program was created that continuously recorded central processing unit (CPU) load averages. …