Flash memory offers attractive features for storage of data, such as non-volatility, shock resistance, fast access speed, and low power consumption. However, it requires erasing before can be overwritten. The erase operations are slow consume comparatively a great deal power. Furthermore, flash only erased limited number times. To overcome hardware limitations, we use the non-in-place update mechanism that cleaner to reclaim space occupied by obsolete data. improve cleaning performance...

This paper describes a flash memory server (FMS) for personal communication devices and embedded home information systems, such as set-top boxes Internet phones. Flash is small, lightweight, shock-resistant, nonvolatile, requires little power. Writing to segments erasing the segment in advance. However, erase operations are slow power-wasting usually decrease system performance. The number of cycles also limited. To reduce needed evenly wear memory, new management scheme has been designed. A...

Embedded systems are usually resource limited in terms of processing power, memory, and power consumption, thus embedded TCP/IP should be designed to make the best use resources. Applying zero-copy mechanism can reduce memory usage CPU time for data transmission. Power consumption reduced as well. In this paper, we present design implementation target component, LyraNET, which is derived from Linux codes remodeled a reusable software component that independent operating hardware. Performance...

Due to the advantages of non-volatility, lightweight, low power consumption, and shock resistance, flash memory has been widely used as storage embedded systems mobile devices. However, unlike hard disk, does not support update-in-place operation, each block on limited erasure cycles. Therefore, needs a different management scheme designed specifically for characteristics. Many researches adopt log-based approach which an efficient cleaning mechanism reclaim space occupied by obsolete data....

Due to the explosive growth of Internet and increasing service demands from all around world, cluster-based system that consists one request-dispatching server several request-handling servers has become a cost-effective way serve huge amount demands. Nowadays, Web have handle more complex types requests since clients may be mixed with dynamic pages, database processing, or multimedia stream data. Therefore, cluster should designed intelligent request dispatching policies for supporting...

Multicore systems with Non-Uniform Memory Access (NUMA) architecture are getting popular in computer systems. Processor cores take longer time to access memories on remote NUMA nodes than local nodes. Because of operating system kernel's memory allocation and load balancing activities, a process may be migrated across its allocated scatter several Remote resource contention cause significant performance degradation multicore In this study, reduce for inter-node interconnect links decrease...

Supporting remote dynamic update on wireless sensor networks (WSNs) allows software of nodes to be dynamically updated or enhanced adapt changed environmental conditions application requirements. Many researches use diff-based approaches transmit only the code difference for updating. For resource-limited WSNs, we have proposed a new approach named Two-Stage Diff components effectively in nodes. By using some carefully designed methods increase similarities between old and program images,...

Abstract To support Web clusters with efficient dispatching mechanisms and policies, we propose a light‐weight TCP connection transfer mechanism, Rebuilding, use it to develop content‐aware request platform, LVS‐CAD, in which the dispatcher can extract analyze content requests then dispatch each by its or type of service requested. efficiently HTTP/1.1 persistent clusters, scheduling should be performed per rather than connection. Consequently, multiple as an extension normal is proposed...

Server clustering is a cost-effective solution to increase the service capacity and system reliability. It also gives greater scalability for handling growing huge amount of demands. Nowadays, cloud platforms take advantage virtualization technology make their actual hosts virtualized. In this study, we explore issues implementing server clusters based on virtual machines (VM), including architectures load distribution algorithms. We utilize Linux Virtual (LVS) design several kinds VM-based...

Summary Although nonuniform memory access architecture provides better scalability for multicore systems, cores accessing on remote nodes take longer than those local nodes. Remote accompanied by contention internode interconnection degrades performance. Properly mapping threads to and data accessed their can substantially improve performance energy efficiency. However, an operating system kernel's load‐balancing activity may migrate across nodes, which thus messes up the thread mapping....

It is always a critical issue for IT managers to balance the servers' load in server cluster. Nowadays, Software-Defined Network (SDN) novel way data centers manage their network environments with better flexibility, reliability, and performance. Therefore, more studies have focused on designed balancing strategies modern environments. In this study, fully utilize hardware resources, we propose Least-Loading (LL) load-balancing strategy based OpenFlow which one of first most widely used SDN...

NUMA multi-core systems divide system resources into several nodes. When an imbalance in the load between cores occurs, kernel scheduler’s balancing mechanism then migrates threads or across Remote memory access is required for a thread to on previous node, which degrades performance. Threads be migrated must selected effectively and efficiently since related operations run critical path of scheduler. This study focuses improving inter-node multithreaded applications. We propose thread-aware...

We have designed and implemented a component-based VFAT file system, LyraFILE. To support the basic operations, we reused some codes from Linux system. In order for it to operate efficiently in resource-limited embedded systems, enhancements are made especially performance improvement memory saving. addition, LyraFILE contains multiple components that can be separately configured or replaced adapting into target environments, changing system specific features, issues. The experience of this...

Log-Structured File System (LFS) is famous for its optimization write performance. Because of append-only nature, garbage collection needed to reclaim the space occupied by obsolete data. The cleaning overhead could significantly decrease performance file system. However, traditional policies do not consider storage location where valid data in cleaned segments should be placed and rewritten to. In this paper, we propose a new method called R-LFS dynamically reorganize disk approximate organ...

With the rapid development of Internet Things (IoT) technology, diversified applications deploy extensive sensors to monitor objects, such as PM2.5 air quality monitoring. The transmit data server periodically and continuously. However, a single cannot provide efficient services for ever-growing IoT devices they generate. This study bases on concept symmetry architecture quantities in system design explores load balancing issue improve performance. uses Linux Virtual Server (LVS)...

Dynamic update provides more flexibility in the development of embedded systems since it allows to their components on-the-fly without rebooting or stopping system services. However, downloading an incautiously developed component may corrupt system. In this paper, we have implemented a platform which supports dynamic dissemination mechanism for upgrading operating whole systems. Besides, resource protection is protect resources. If error has misused resources, wasted resources will be...