• The KIPS Transactions:PartB
• /
• v.12B no.7 s.103
• /
• pp.775-782
• /
• 2005

In this paper, we propose multicast technique in order to reduce the required network bandwidth by n times, by merging the adjacent multicasts depending on the number of HENs (Head-End-Nodes) n that request the same video. Allowing new clients to immediately join an existing multicast through patching improves the efficiency of the multicast and offers services without any initial latency. A client might have to download data through two channels simultaneously, one for multicast and the other for patching. The more the frequency of requesting the video is, the higher the probability of caching it among HENs increases. Therefore, the requests for the cached video data can be served by HENs. Multicast from server is generated when the playback time exceeds the amount of cached video data. Since the interval of multicast can be dynamically expanded according to the popularity of videos, it can be reduced the server's workload and the network bandwidth. We perform simulations to compare its performance with that of conventional multicast. From simulation results, we confirm that the Proposed multicast technique offers substantially better performance.

• KIISE Transactions on Computing Practices
• /
• v.21 no.8
• /
• pp.543-548
• /
• 2015

Recently, volume of data has been growing dramatically along with the growth of social media and digital devices. However, the existing disk-based distributed file systems have limits to their performance of data processing or data access, due to I/O processing costs and bottlenecks. To solve this problem, the caching technique is being used to manage data in the memory. In this paper, we propose a cache management scheme to handle load balancing in a distributed memory environment. The proposed scheme distributes the data according to the memory size, n distributed environments with different memory sizes. If overloaded nodes occur, it redistributes the the access time of the caching data. In order to show the superiority of the proposed scheme, we compare it with an existing distributed cache management scheme through performance evaluation.

• Journal of KIISE:Databases
• /
• v.30 no.4
• /
• pp.408-419
• /
• 2003

Data warehouses are usually dedicated to the processing of quires issued by decision support system(DSS). The response time of DSS queries is typically several orders of magnitude higher than the one of OLTP queries. Since DSS queries are often submitted interactively, techniques for reducing their response time are important. The caching of query results is one such technique particularly well suited to the DSS environment. In this paper, we present a cache manager for such an environment. Specifically, we define a canonical form of query. The cache manager looks up a query based on the exact query match or using a suggested query split process if the query is found is non-canonical form or in canonical form, respectively. It dynamically maintains the cache content by employing a profit function which reflects in an integrated manner the query execution cost, the size of query result, the reference rate, the maintenance cost of each result due to updates of their base tables, and the frequency of such updates. We performed the experimental evaluation and it positively shows the performance benefit of our cache manager.

• The KIPS Transactions:PartA
• /
• v.8A no.4
• /
• pp.447-454
• /
• 2001

Various caching and prefetching algorithms have been investigated to identify and effective method for improving the performance of I/O devices. A prefetching algorithm decreases the processing time of a system by reducing the number of disk accesses when an I/O is needed. This paper proposes an AMBA prefetching method that is an extended version of the OBA prefetching method. The AMBA prefetching method will prefetching blocks continuously as long as disk bandwidth is enough. In this method, though there were excessive data request rate, we would expect efficient prefetching. And in the AMBA prefetching method, to prevent the cache pollution, it limits the number of data blocks to be prefetched within the cache area. It can be implemented in a user-level File System based on a Linux Operating System. In particular, the proposed prefetching policy improves the system performance by about 30∼40% for large files that are accessed sequentially.

• Journal of KIISE
• /
• v.42 no.9
• /
• pp.1109-1116
• /
• 2015

In this paper, we propose a query processing approach that uses the Spark functional programming and distributed memory system to solve the computational overhead of SPARQL. In the semantic web, RDF ontology data is produced at large scale, and the main challenge for the semantic web is to query and manipulate such a large ontology with a high throughput. The most existing studies on SPARQL have focused on deploying the Hadoop MapReduce framework, and although approaches based on Hadoop MapReduce have shown promising results, they achieve a low level of throughput due to the underlying distributed file processes. Therefore, in order to speed up the query processes, we suggest query- processing methods that are based on memory caching in distributed memory system. Our approach is also integrated with a clause unification method for propagating between the clauses that exploits Spark join, map and filter methods along with caching. In our experiments, we have achieved a high level of performance relative to other approaches. In particular, our performance was nearly similar to that of Sempala, which has been considered to be the fastest query processing system.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.5
• /
• pp.37-44
• /
• 2009

Recently, as the quality of multimedia data gets higher, multimedia servers require larger storage capacity and higher I/O bandwidth. In these large scale multimedia servers, the load-unbalance problem among disks due to the difference in access frequencies to multimedia objects according to their popularities significantly affects the system performance. To address this problem, many data replication schemes have been proposed. In this paper, we propose a novel data migration/replication scheme to provide better storage efficiency and performance than the dynamic data replication scheme which is typical data replication scheme employed in multimedia servers. This scheme can reduce the additional storage space required for replication, which is a major defect of replication schemes, by decreasing the number of copies per object. The scheme can also increase the number of concurrent users by increasing the caching effect due to the reduced lengths of the intervals among requests for each object.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.2
• /
• pp.93-98
• /
• 2024

With the rapid advance in AI (artificial intelligence) and high-performance computing technologies, deep learning is being used in various fields. Deep learning proceeds training by randomly reading a large amount of data and repeats this process. A large number of files are randomly repeatedly referenced during deep learning, which shows different access characteristics from traditional workloads with temporal locality. In order to cope with the difficulty in caching caused by deep learning, we propose a new sampling method that aims at reducing the randomness of dataset reading and adaptively operating on existing buffer cache algorithms. We show that the proposed policy reduces the miss rate of the buffer cache by 16% on average and up to 33% compared to the existing method, and improves the execution time by up to 24%.

• The Journal of Society for e-Business Studies
• /
• v.11 no.1
• /
• pp.1-23
• /
• 2006

As data on the Web is more and more in XML due to proliferation of Web applications such as e-Commerce, it is strongly required to rapidly process XML queries. One of such techniques is XML query caching. For frequently submitted queries, their results could be cached in order to guarantee fast response for the same queries. In this paper, we propose techniques for XML query performance improvement whereby the set of node identifiers(NIS) for an XML query is cached. NIS is most commonly employed as a format of XML query result,, consisting of the identifiers of the XML elements that comprise the query result. With NIS, it is suitable to meet the Web applications data retrieval requirements because reconstruction and/or modification of query results and integration of multiple query results can be efficiently done. Incremental refresh of NIS against its source updates can also be efficiently done. When the query result is requested in XML, however, materialization of NIS is needed by retrieving the source XML elements through their identifiers. In this paper, we consider three different types of NISs. proposing the algorithms of their creation, materialization, and incremental refresh. All of them were implemented using an RDBMS. Through a detailed set of performance experiments, we showed the efficiency of the proposed XML query caching techniques.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.11
• /
• pp.3116-3123
• /
• 1999

Current disk systems are generally designed to reduce read traffic more effectively. Hence, write traffic of the I/O workload could potentially become a bottleneck of the disk system performance. In order to overcome this problem without much cost, this paper presents using outermost-zone track of multi-zoned recording disk as a secondary disk cache. The proposed disk cache improves the disk system performance by following exploitations: speed difference between block transfer and track transfer, difference in transfer rate between outermost-zone tracks and inner tracks, reduction in the seek time caused by decreasing the number of disk cache tracks, and idle period during burst accesses. In addition, it does not waste the disk space because it allocates the caching space by the cylinder unit. The simulation results show that the proposed system improves 2.54∼3.11 times better in terms of average response time for write operations than existing disk systems..

• Journal of information and communication convergence engineering
• /
• v.11 no.3
• /
• pp.173-178
• /
• 2013

To discover objects of interest in unstructured peer-to-peer networks, the peers rely on flooding query messages which create incredible network traffic. This article evaluates the performance of an unstructured Gnutella-like protocol over mobile ad-hoc networks and proposes modifications to improve its performance. This paper offers an enhanced mechanism for an unstructured Gnutella-like network with improved peer features to better meet the mobility requirement of ad-hoc networks. The proposed system introduces a novel caching optimization technique and enhanced ultrapeer selection scheme to make communication more efficient between peers and ultrapeers. The paper also describes an enhanced query mechanism for efficient searching by applying multiple walker random walks with a jump and replication technique. According to the simulation results, the proposed system yields better performance than Gnutella, XL-Gnutella, and random walk in terms of the query success rate, query response time, network load, and overhead.