• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.5
• /
• pp.282-296
• /
• 2006

In a wireless client server environment, data broadcasting is an efficient data dissemination method; a server broadcasts data, and some of broadcasted data are cached in a mobile client's cache to save the narrow communication bandwidth, limited resources, and data access time. A server also broadcasts invalidation reports to maintain the consistency between server data and a client's cached data. Most of existing works on the cache consistency problems simply purge the entire cache when the disconnection time is long enough to miss the certain amount(window size) of IRs. This paper presents a cache invalidation method to increase mobile clients' cache reusability in case of a long disconnection. Instead of simply dropping the entire cache regardless of its consistency, a client estimates the cost of purging all the data with the cost of selective purge. If the cost of dropping entire cache is higher, a client maintains the cache and selectively purge inconsistent data using uplink bandwidth for validation request. The simulation results show that this scheme increases the cache reusability since it effectively considers the update rates and the broadcast frequencies of cached data in estimating the cost of cache maintenance.

• Journal of KIISE:Databases
• /
• v.34 no.3
• /
• pp.217-231
• /
• 2007

Frequent disconnection is connected directly to client's cache consistency problem in Mobile Computing environment. For solving cache consistency problem, research about Invalidation Report is studied. But, existent invalidation report structure comes with increase of size of invalidation report structure and decline of cache efficiency if quantity of data become much, or quantity of updated data increases. Also, while existent method confirms whole cache, invalidation report doesn't support selective listening. This paper proposes TTCI(Timestamp Tree-based Cache Invalidation scheme) as invalidation report structure that solve problem of these existing schemes and improve efficiency. We can make TTCI using timestamp of updated data, composing timestamp tree and list ID of data in updated order. If we utilize this, each client can confirm correct information in point that become own disconnecting and increase cache utilization ratio. Also, we can pare down client's resources consumption by selective listening using tree structure. We experimented in comparison with DRCI(Dual-Report Cache Invalidation) that is existent techniques to verify such efficiency of TTCI scheme.

• Journal of IKEEE
• /
• v.8 no.2 s.15
• /
• pp.207-215
• /
• 2004

In a hierarchical-bus multiprocessor system, cache coherence protocol has effect on system performance. Under a particular cache coherence protocol, system performance can be affected by bus bandwidth, memory size, and memory block size. Therefore sensitivity analysis is necessary for the part of multiprocessor system. In this paper, we set up cache coherence protocol for hierarchical-bus multiprocessor system, and compute probability of state of protocol, and analyze sensitivity for part of system by simulation.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10c
• /
• pp.9-11
• /
• 2000

최근의 네트웍 환경에서는 멀티미디어 서비스와 대용량의 파일을 이용하는 어플리케이션의 증가로 인해 이를 충족시킬 수 있는 저장 장치가 요구되고 있는 실정이며 이러한 저장 장치를 이용한 분산 환경의 네트웍 파일 시스템이 필수적이 되었다. 실제로 ATM, Fast switched LAN, Fibre channel 같은 고속의 네트웍 발달로 인해 분산 환경의 네트웍 파일 시스템에서 디스크를 엑세스하는 속도보다 오히려 네트웍으로 연결된 원격지의 메모리를 액세스하는 것이 더 빨라졌다. 따라서 지역 디스크 캐싱 기법이 분산 환경의 네트웍 저장 시스템으로 적용되면서 전역적 메모리를 관리하는 것과 원격지간의 캐시 일관성 문제(cache consistency)를 고려하지 않을 수 없게 되었다. 본 논문에서는 분산 환경의 캐싱 기법을 살펴보고 전역적 메모리의 캐시 일관성 문제를 다루면서 이에 대한 설계방안 및 앞으로의 연구 방향을 제시한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10c
• /
• pp.513-515
• /
• 1999

기하급수적으로 증가하는 인터넷 트래픽(traffic)의 대처방안으로 캐시 일관성 유지, 캐싱 알고리즘 등 웹 캐싱에 대한 연구가 끊임없이 진행되고 있다. 본 논문에서는 인터넷 트래픽의 양을 줄이기 위해 다중의 캐시 서버를 설계하였다. 다중의 캐시 서버는 단일 캐시 서버를 사용할 때 발생하는 웹 서버의 부하를 감소시켜 네트웍 트래픽의 양을 줄여준다. 또한 다중의 캐시는 라우터, 디스크, 메모리 같은 하드웨어의 비용을 절감할 수 있어 경제적인 측면에서도 효율적이다.

• Journal of KIISE:Databases
• /
• v.32 no.3
• /
• pp.276-284
• /
• 2005

In the mobile computing environment, caching data at the client's side is a useful technique for improving the performance. For caching to be effective, the cache contents must be consistent with those stored in the server. That is, unfortunately, difficult to enforce due to the frequent disconnection and mobility of clients. In the literature, the basic approach adopted is for the sever to periodically broadcast invalidation reports(IRs) that contain information about objects that have been updated recently. However IR-based cache maintenance scheme has a main drawback. Any client who has been disconnected longer than w seconds has to discard all cached items even some of them may still be valid. In this paper, we propose two main schemes which can adapt clients' losing IR messages by long disconnection. First, the client sends its disconnection time when connection is re-established, and the server uses this information to decide the size of broadcast window w dynamically. Second, the server maintains IR messages for a given period of time and a client who has been disconnected longer than w seconds can verify its cache consistency individually. Compared to previous IR-based schemes, our schemes can significantly reduce cache missing and uplink request and improve the throughput by reducing response time.

• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1245-1251
• /
• 2017

In this study we present a novel directory architecture that can dynamically allocate a directory entry for a cache block on demand at runtime only when the block is shared by more than one core. Thus, we do not maintain coherence for private blocks, substantially reducing the number of directory entries. Even for shared blocks, we allocate directory entry dynamically only when the block is actively shared, further reducing the number of directory entries at runtime. For this, we propose a new directory architecture called dynamic directory table (DDT), which is implemented as a cache of active directory entries. Through our detailed simulation on PARSEC benchmarks, we show that DDT can outperform the expensive full-map directory by a slight margin with only 17.84% of directory area across a variety of different workloads. This is achieved by its faster access and high hit rates in the small directory. In addition, we demonstrate that even smaller DDTs can give comparable or higher performance compared to recent directory optimization schemes such as SPACE and DGD with considerably less area.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.9
• /
• pp.699-708
• /
• 2006

Bus splitting technique reduces bus energy by placing modules with frequent communications closely and using necessary bus segments in communications. But, previous bus splitting techniques can not be used in MPSoC platform, because it uses cache coherency protocol and all processors should be able to see the bus transactions. In this paper, we propose a bus splitting technique for MPSoC platform to reduce bus energy. The proposed technique divides a bus into several bus segments, some for private memory and others for shared memory. So, it minimizes the bus energy consumed in private memory accesses without producing cache coherency problem. We also propose a task allocation technique considering cache coherency protocol. It allocates tasks into processors according to the numbers of bus transactions and cache coherence protocol, and reduces the bus energy consumption during shared memory references. The experimental results from simulations say the bus splitting technique reduces maximal 83% of the bus energy consumption by private memory accesses. Also they show the task allocation technique reduces maximal 30% of bus energy consumed in shared memory references. We can expect the bus splitting technique and the task allocation technique can be used in multiprocessor platforms to reduce bus energy without interference with cache coherency protocol.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.10
• /
• pp.13-21
• /
• 2013

Recently, although the number of cores on a chip multi-processor increases, multi-programming or multi-threaded programming techniques to utilize the whole cores are still insufficient. Therefore, there inevitably exist some idle cores which are not working. This results in a waste of the caches, so-called idle caches which are dedicated to those idle cores. In this research, we propose amethodology to exploit idle caches effectively as victimcaches of on-chip memory resource. In simulation results, we have achieved 19.4%and 10.2%IPC improvement in 4-core and 16-core respectively, compared to previous technique.

• Journal of KIISE:Databases
• /
• v.28 no.2
• /
• pp.253-265
• /
• 2001

본 논문은 이동 컴퓨팅 환경에서 연속 부분 질의의 효과적인 처리를 위한 캐시 관리 방안을 제안한다. 연속 부분 부합 질의란 질의의 결과가 클라이언트의 메모리에 일관성을 유지하면서 지속되는 부분 부합 질의이다. 기존의 이동 환경을 위한 캐시 관리 기법은 레코드 식별자를 기반으로 하는 방법들이다. 하지만, 부분 부합 질의는 데이터의 내용을 기반으로 탐색하는 질의이기 때문에 이러한 레코드 식별자를 기반으로 하는 방법들은 캐시 관리를 효율적으로 할 수 없다. 제안하는 캐시 관리 방안에서는, 이동 클라이언트의 캐시 상태를 프레디킷(predicate)으로 기술하고, 서버가 캐시 관리를 위해 클라이언트에게 방송하는 캐시 무효화 정보, 즉 Cache Invalidation Reports(CIR)을 프레디킷으로 구성한다. 이러한 프레디킷 표현을 사용하여, 일련의 캐시 관리 기법-억지 방법(the brute-force method), 빼기 방법(the subtraction method), 교차 방법(the intersection method)-들을 제안한다. 그리고, 제안하는 방법의 계산 복잡도를 계산한다.