Dongwon Kang

μ-tree: an ordered index structure for NAND flash memory

As NAND flash memory becomes increasingly popular as data storage for embedded systems, many file systems and database management systems are being built on it. They require an efficient index structure to locate a particular item quickly from a huge amount of directory entries or database records. This paper proposes μ-Tree, a new ordered index structure tailored to the characteristics of NAND flash memory. μ-Tree is a balanced tree similar to B+-Tree. In μ-Tree, however, all the nodes along the path from the root to the leaf are put together into a single flash memory page in order to minimize the number of flash write operations when a leaf node is updated. Our experimental evaluation shows that μ-Tree outperforms B+-Tree by up to 28% for traces extracted from real workloads. With a small in-memory cache of 8 Kbytes, μ-Tree improves the overall performance by up to 90% compared to B+-Tree with the same cache size.As NAND flash memory becomes increasingly popular as data storage for embedded systems, many file systems and database management systems are being built on it. They require an efficient index structure to locate a particular item quickly from a huge amount of directory entries or database records. This paper proposes μ-Tree, a new ordered index structure tailored to the characteristics of NAND flash memory. μ-Tree is a balanced tree similar to B+-Tree. In μ-Tree, however, all the nodes along the path from the root to the leaf are put together into a single flash memory page in order to minimize the number of flash write operations when a leaf node is updated. Our experimental evaluation shows that μ-Tree outperforms B+-Tree by up to 28% for traces extracted from real workloads. With a small in-memory cache of 8 Kbytes, μ-Tree improves the overall performance by up to 90% compared to B+-Tree with the same cache size.

μ-FTL:: a memory-efficient flash translation layer supporting multiple mapping granularities

NAND flash memory is being widely adopted as a storage medium for embedded devices. FTL (Flash Translation Layer) is one of the most essential software components in NAND flash-based embedded devices as it allows to use legacy files systems by emulating the traditional block device interface on top of NAND flash memory. In this paper, we propose a novel FTL, called μ-FTL. The main design goal of μ-FTL is to reduce the memory foot-print as small as possible, while providing the best performance by supporting multiple mapping granularities based on variable-sized extents. The mapping information is managed by μ-Tree, which offers an efficient index structure for NAND flash memory. Our evaluation results show that μ-FTL significantly outperforms other block-mapped FTLs with the same memory size by up to 89.7%.

Constraint-based human resource allocation in software projects

Resource allocation in a software project is crucial for successful software development. Among various types of resources, human resource is the most important as software development is a human‐intensive activity. Human resource allocation is very complex owing to the human characteristics of developers. The human characteristics affecting allocation can be grouped into individual‐level characteristics and team‐level characteristics. At the individual level, familiarity with tasks needs to be taken into account as it affects the performance of developers. In addition, developers have different levels of productivity, depending on their capability and experience; the productivity of developers also varies according to tasks. At the team level, characteristics such as team cohesion, communication overhead, and collaboration and management also affect human resource allocation. As these characteristics affect the efficiency of project execution, we treat them as constraints of human resource allocation in our approach. We identify individual‐level constraints and team‐level constraints based on the literature and interviews with experts in the industry. With these constraints, our approach optimizes the scheduling of human resource allocations, resulting in more realistic and efficient allocations. We also provide a guideline supporting various factors, with respect to roles and module characteristics, to estimate the productivity of developers based on COCOMO II. As productivity data are hard to obtain and manage, our guideline can provide a useful direction for human resource allocation in case of software projects. To validate our proposed approach, we document a case study using real project data. Copyright

A Case Retrieval Method for Knowledge-Based Software Process Tailoring Using Structural Similarity

Reuse of the software development process and its knowledge and experiences is a critical factor for the success of the software project. On the other hand, the software development process needs to be tailored to reflect the specific characteristics of the software project. So, if we can retrieve a similar process to a new project, process tailoring will be less costly and less error-prone because the retrieved process can be tailored to the new case with fewer modifications. In this paper, we propose the case retrieval method based on structural similarity. The structural similarity is calculated by the degree that process elements in a past case are applicable to a new project. By measuring the structural similarity, the retrieved process is ensured to be tailored to the new case with fewer modifications. We validate the usefulness of our method through the experiments using 30 cases.

μ*-Tree: An Ordered Index Structure for NAND Flash Memory with Adaptive Page Layout Scheme

As NAND flash memory is gaining popularity as a storage medium for mobile embedded devices, many flash-aware file systems, flash-aware DBMSes, and flash translation layers (FTLs) require an flash-efficient index structure. This paper proposes a novel index structure called μ*-Tree which natively works on NAND flash memory, aiming at improving performance over B+-Tree. μ*-Tree stores all the nodes along the path from the root to the leaf into a single flash memory page in order to minimize the number of flash write operation when a node is updated. Furthermore, μ*-Tree has an adaptive page layout scheme which dynamically adjusts the page layout according to the workload characteristics on-the-fly. μ*-Tree also allows flash pages with different page layouts to coexist in the same tree. Our evaluation results with real workload traces show that μ*-Tree outperforms B+-Tree by up to 55 percent in terms of the time needed for flash operations. With a small in-memory cache of 32 KB, μ*-Tree improves the overall performance by up to five times compared to B+-Tree with the same cache size.

Developing a software process simulation model using SPEM and analytical models

To develop Software Process Simulation Models (SPSMs) more effectively, we need to consider the following issues: reducing the knowledge difference between increasing the reusability of a simulation model, decreasing the complexity of a simulation model, and mitigating the lack of historical data. We propose an approach to develop an SPSM based on SPEM, and widely adopted analytical models. An SPEM-based process model is integrated with the quantitative information and transformed into a DEVS-Hybrid SPSM. Our approach resolves the issues by the transformation algorithms, the hierarchical and modularised modelling properties of SPEM and DEVS, and widely adopted analytical models.

Software Fault Prediction Models for Web Applications

Our daily life increasingly relies on Web applications. Web applications provide us with abundant services to support our everyday activities. As a result, quality assurance for Web applications is becoming important and has gained much attention from software engineering community. In recent years, in order to enhance software quality, many software fault prediction models have been constructed to predict which software modules are likely to be faulty during operations. Such models can be utilized to raise the effectiveness of software testing activities and reduce project risks. Although current fault prediction models can be applied to predict faulty modules of Web applications, one limitation of them is that they do not consider particular characteristics of Web applications. In this paper, we try to build fault prediction models aiming for Web applications after analyzing major characteristics which may impact on their quality. The experimental study shows that our approach achieves very promising results.

Developing a Simulation Model Using a SPEM-Based Process Model and Analytical Models

It is hard to adopt a simulation technology because of the difficulty in developing a simulation model. In order to resolve the difficulty, we consider the following issues: reducing the cost to develop a simulation model, reducing the simulation model complexity, and resolving the lack of historical data. We propose an approach to deriving a simulation model from a descriptive process model and widely adopted analytical models. We provide a method to develop simulation models and a tool environment to support the method. We applied our approach in developing the simulation model for a government project. Our approach resolves the issues by the transformation algorithms, the hierarchical and modularized modeling properties of UML and (Discrete Event System Specification) DEVS, and widely adopted analytical models.