Gerald Parks Bozman

Analysis of free-storage algorithms: revisited

Most research in free-storage management has centered around strategies that search a linked list and strategies that partition storage into predetermined sizes. Such algorithms are analyzed in terms of CPU efficiency and storage efficiency. The subject of this study is the free-storage management in the Virtual Machine/System Product (VM/SP) system control program. As a part of this study, simulations were done of established, and proposed, dynamic storage algorithms for the VM/SP operating system. Empirical evidence is given that simplifying statistical assumptions about the distribution of interarrival times and holding times has high predictive ability. Algorithms such as first-fit, modified first-fit, and best-fit are found to be CPU-inefficient. Buddy systems are found to be very fast but suffer from a high degree of internal fragmentation. A form of extended subpooling is shown to be as fast as buddy systems with improved storage efficiency. This algorithm was implemented for VM/SP, and then measured. Results for this algorithm are given for several production VM/SP systems.

The software lookaside buffler reduces search overhead with linked lists

Linked lists are one of the most f requent ly used data s t ructures in compu te r sof tware [3, 5]. Unfor tuna te ly thei r advantages are often offset by a high search overhead. This is especial ly t rue in mach ines that have a cache I [6] because the dis tance b e t w e e n list e lements often exceeds the cache line, that is, the q u a n t u m of storage in the cache. On the IBM 3081, for example , a cache miss can requi re 20-30 t imes more mach ine cycles than a cache hit. Since it is not unusua l for the search dis t r ibut ion against l inked lists to be ex t r eme ly nonuni form, a software lookaside buffer 1 can yield significant performance benefi ts at the expense of a modes t analysis and p rogramming effort. In this context , a lookaside buffer is the software analog to the ha rdware cache in that a hit in the lookaside buffer can save significant overhead in accessing an i tem in the l inked list. Al though lookaside buffers can be appl ied to o ther data structures, we have seen the greatest ove rhead reduc t ion wi th l inked lists.

VM/XA SP2 minidisk cache

Given the growing disparity between CPU power and the speed of secondary storage, a data cache exploiting large processor storage has the potential to improve response time dramatically in many situations. The VM/XA SP2 minidisk cache facility, the result of research activity on the characteristics of interactive file-system activity, uses expanded storage to cache input/output to minidisks on the Conversational Monitor System. The size of the cache is dynamically adjusted by an arbitration process to optimize system performance. Several other functions improve the performance of the cache during periods of unusual I/O loads.

Multimedia file serving with the OS/390 LAN server

The rapidly increasing storage and transmission capacities of computers and the progress in compression algorithms make it possible to build multimedia applications that include audio and video. Such applications range from educational and training videos, delivered to desktops in schools and enterprises, to entertainment services at home. Applications developed for stand-alone personal computers can be deployed in distributed systems without change by using the client/server model and file servers that allow the sharing of applications among many users. The OS/390™ LAN Server has been enhanced to support multimedia data delivery. Resource management and admission control, wide disk striping to provide high data bandwidths, and multimedia-specific performance enhancements have been added. The resulting server benefits from the robustness, scalability, and flexibility of the S/390® system environment, which allows it to move into new multimedia applications. Multimedia support on a robust, widely installed platform with little or no additional hardware requirements gives customers the opportunity to enhance their existing applications with multimedia features and then expand their capacity as the demands of the applications increase. This multimedia server platform is in use with several interesting applications.

The parallel processing compute server

The Parallel Processing Compute Server (PPCS) is a distributed-memory multiprocessing system consisting of System/370 microprocessors (33 at present) interconnected through a matrix switch. This paper describes the hardware configuration, the extensions to the System/370 instruction set that are provided to support the distributed memory and interprocessor signalling, the modification to the VM/SP operating system that allow it to run effectively on many closely coupled processors (most of which have no disks), and the application-support layer, which permits FORTRAN programs to take advantage of the highly parallel environment. A PPCS with 32 satellite processors, Parallel VM, and CS/X has been installed at the European Organization for Nuclear Research (CERN) and is running applications used by the high-energy physics (HEP) community. At CERN, on-line data acquisition and preprocessing of experimental data is a major computing task. High-energy physics experiments can generate streams of data at a rate of several gigabytes per second. Once filtered and recorded, data are subject to further analysis off-line. Both on-line and off-line processing are numerically intensive computing jobs. The PPCS is primarily intended for the off-line processing, but it can be used for on-line work as well.