Rights to individual papers remain with the author or the author's employer. Permission is granted for noncommercial reproduction of the work for educational or research purposes. This copyright notice must be included in the reproduced paper. USENIX acknowledges all trademarks herein.

AbstractÐEfficient and effective buffering of disk blocks in main memory is critical for better file system performance due to a wide speed gap between main memory and hard disks. In such a buffering system, one of the most important design decisions is the block replacement policy that determines which disk block to replace when the buffer is full. In this paper, we show that there exists a spectrum of block replacement policies that subsumes the two seemingly unrelated and independent Least Recently Used (LRU) and Least Frequently Used (LFU) policies. The spectrum is called the LRFU (Least Recently/Frequently Used) policy and is formed by how much more weight we give to the recent history than to the older history. We also show that there is a spectrum of implementations of the LRFU that again subsumes the LRU and LFU implementations. This spectrum is again dictated by how much weight is given to recent and older histories and the time complexity of the implementations lies between O(1) (the time complexity of LRU) and O…log 2 n† (the time complexity of LFU), where n is the number of blocks in the buffer. Experimental results from trace-driven simulations show that the performance of the LRFU is at least competitive with that of previously known policies for the workloads we considered. Index TermsÐBuffer cache, LFU, LRU, replacement policy, trace-driven simulation. 1

In traditional file system implementations, the Least Recently Used (LRU) block replacement scheme is widely used to manage the buffer cache due to its simplicity and adaptability. However, the LRU scheme exhibits performance degradations because it does not make use of reference regularities such as sequential and looping references. In this paper, we present a Unified Buffer Management (UBM) scheme that exploits these regularities and yet, is simple to deploy. The UBM scheme automatically detects sequential and looping references and stores the detected blocks in separate partitions of the buffer cache. These partitions are managed by appropriate replacement schemes based on their detected patterns. The allocation problem among the divided partitions is also tackled with the use of the notion of marginal gains. In both trace-driven simulation experiments and experimental studies using an actual implementation in the FreeBSD operating system, the performance gains obtained through the use of this scheme are substantial. The results show that the hit ratios improve by as much as 57.7 % (with an average of 29.2%) and the elapsed times are reduced by as much as 67.2 % (with an average of 28.7%) compared to the LRU scheme for the workloads we used.

Although today's computers provide huge amounts of main memory, the ever-increasing load of large data servers, imposed by resource-intensive decision-support queries and accesses to multimedia and other complex data, often leads to memory contention and may result in severe performance degradation. Therefore, careful tuning of memory mangement is crucial for heavy-load data servers. This paper gives an overview of self-tuning methods for a spectrum of memory management issues, ranging from traditional caching to exploiting distributed memory in a server cluster and speculative prefetching in a Web-based system. The common, fundamental elements in these methods include on-line load tracking, near-future access prediction based on stochastic models and the available on-line statistics, and dynamic and automatic adjustment of control parameters in a feedback loop. 1 The Need for Memory Tuning Although memory is relatively inexpensive and modern computer systems are amply equipp...

Two contributions are made in this paper. First, we show that system level characterization of file block references is inadequate for maximizing buffer cache performance. We show that a finer-grained characterization approach is needed. Though application level characterization methods have been proposed, this is the first attempt, to the best of our knowledge, to consider file level characterizations. We propose an Application/File-level Characterization (AFC) scheme where we detect on-line the reference characteristics at the application level and then at the file level, if necessary. The results of this characterization are used to employ appropriate replacement policies in the buffer cache to maximize performance. The second contribution is in proposing an efficient and fair buffer allocation scheme. Application or file level resource management is infeasible unless there exists an allocation scheme that is efficient and fair. We propose the \DeltaH IT allocation scheme that tak...

In this paper, we propose a new adaptive buffer management scheme called DEAR (DEtection based Adaptive Replacement) that automatically detects the block reference patterns of applications and applies different replacement policies to different applications based on the detected reference pattern. The proposed DEAR scheme uses a periodic process. Detection is made by associating block attribute values such as backward distance and frequency gathered at the (i \Gamma 1)-th invocation with forward distances of blocks referenced between the (i \Gamma 1)-th and i-th invocations. We implemented the DEAR scheme in FreeBSD 2.2.5 and measured its performance using several real applications. The results show that compared with the LRU buffer management scheme, the proposed scheme reduces the number of disk I/Os by up to 51% (with an average of 23%) and the response time by up to 35% (with an average of 12%) in the case of single application executions. For multiple applications, the proposed sc...

Recent research has shown that near optimal performance can be achieved by adaptive block replacement policies that use user-level hints regarding the block reference pattern. However, obtaining user-level hints requires considerable effort from users making it difficult to apply adaptive replacement policies to diverse kinds of applications. We propose a new adaptive block management scheme that we call DEAR (DEtection based Adaptive Replacement) which makes on-line detections of block reference patterns of applications using Decision Trees without user intervention. Based on the detected reference pattern, DEAR applies an appropriate replacement policy to each application. This scheme is suitable for buffer management in systems such as multimedia servers where data reference patterns of applications may be diverse. Results from trace driven simulations show that the DEAR scheme can detect the reference patterns of applications and reduce the miss ratio up to 15 percentage points com...

A new buffer replacement scheme called DEAR (DEtection-based Adaptive Replacement) is presented for effective caching of disk blocks in the operating system. The proposed DEAR scheme automatically detects block reference patterns of applications and applies different replacement policies to different applications depending on the detected reference pattern. The detection is made by a periodic process and is based on the relationship between block attribute values such as backward distance and frequency gathered in a period and the forward distance observed in the next period. This paper also describes an implementation and performance measurement of the DEAR scheme in FreeBSD. The results from performance measurements of several real applications show that compared with the LRU scheme, the proposed scheme reduces the number of disk I/Os by up to 51% (with an average of 23%) and the response time by up to 35% (with an average of 12%) in the case of single application execution...

Virtual memory tiling enables applications efficiently to handle much larger arrays of spatial data than is otherwise possible, without requiring expensive additional computing resources. It has particular application to geographical information systems (GIS) with the wide availability of large sets of digital spatial data from remote sensing and other sources. The size of these data sets often greatly exceeds the capabilities of most applications on standard computer platforms. In this paper a virtual memory tiling approach is developed and implemented in the C++ language. A tiled array class with a similar syntax and usage to standard arrays is constructed which is readily integrated with existing algorithms and applications. A framework is developed for classifying operations on spatial data in terms of small and large regions. These two categories are representative of a broad range of operations on spatial data in GIS. Results show that with current levels of processing power and disk storage, processing of arrays in the range of hundreds of megabytes and above is possible at present on desktop platforms.

To overcome the speed gap between processors and disks, many computer systems utilize buffer cache located in main memory. Previous works on buffer cache management mainly use a single fixed block replacement policy such as LRU policy. However, block reference patterns of applications vary depending on the nature of applications, which requires applying different replacement policies to different applications. In this paper, we propose a new adaptive buffer management scheme. Our approach is based on the automatic detection of block reference patterns of applications. The detection is made by associating block attributes such as backward distance and frequency with forward distance of a block. The scheme can classifies four reference patterns: sequential, looping, temporally-clustered, and probabilistic. According to the detected patterns, the scheme applies an appropriate block replacement policy to each application. Also, the scheme employs an efficient block allocation strategy to m...