Soyoon Lee

CLOCK-DWF: A Write-History-Aware Page Replacement Algorithm for Hybrid PCM and DRAM Memory Architectures

Phase change memory (PCM) has emerged as one of the most promising technologies to incorporate into the memory hierarchy of future computer systems. However, PCM has two critical weaknesses to substitute DRAM memory in its entirety. First, the number of write operations allowed to each PCM cell is limited. Second, write access time of PCM is about 6-10 times slower than that of DRAM. To cope with this situation, hybrid memory architectures that use a small amount of DRAM together with PCM have been suggested. In this paper, we present a new memory management technique for hybrid PCM and DRAM memory architecture that efficiently hides the slow write performance of PCM. Specifically, we aim to estimate future write references accurately and then absorb frequent memory writes into DRAM. To do this, we analyze the characteristics of memory write references and find two noticeable phenomena. First, using write history alone performs better than using both read and write history in estimating future write references. Second, the frequency characteristic is a better estimator than temporal locality in predicting future memory writes. Based on these two observations, we present a new page replacement algorithm called CLOCK-DWF (CLOCK with Dirty bits and Write Frequency) that significantly reduces the number of write operations that occur on PCM and also increases the lifespan of PCM memory.

Characterizing Memory Write References for Efficient Management of Hybrid PCM and DRAM Memory

In order to reduce the energy dissipation in main memory of computer systems, phase change memory (PCM) has emerged as one of the most promising technologies to incorporate into the memory hierarchy. However, PCM has two critical weaknesses to substitute DRAM memory in its entirety. First, the number of write operations allowed to each PCM cell is limited. Second, write access time of PCM is about 6-10 times slower than that of DRAM. To cope with this situation, hybrid memory architectures that use a small amount of DRAM together with PCM memory have been suggested. In this paper, we present a new memory management technique for hybrid PCM and DRAM memory architecture that efficiently hides the slow write performance of PCM. Specifically, we aim to estimate future write references accurately and then absorb most memory writes into DRAM. To do this, we analyze the characteristics of memory write references and find two noticeable phenomena. First, using write history alone performs better than using both read and write history in estimating future write references. Second, the frequency characteristic is a better estimator than temporal locality but combining these two properties appropriately leads to even better results. Based on these two observations, we present a new page replacement algorithm called CLOCK-DWF (CLOCK with Dirty bits and Write Frequency) that significantly reduces the number of write operations that occur on PCM.

P/PA-SPTF: Parallelism-aware request scheduling algorithms for MEMS-based storage devices

MEMS-based storage is foreseen as a promising storage media that provides high-bandwidth, low-power consumption, high-density, and low cost. Due to these versatile features, MEMS storage is anticipated to be used for a wide range of applications from storage for small handheld devices to high capacity mass storage servers. However, MEMS storage has vastly different physical characteristics compared to a traditional disk. First, MEMS storage has thousands of heads that can be activated simultaneously. Second, the media of MEMS storage is a square structure which is different from the platter structure of disks. This article presents a new request scheduling algorithm for MEMS storage called P-SPTF that makes use of the aforementioned characteristics. P-SPTF considers the parallelism of MEMS storage as well as the seek time of requests on the two dimensional square structure. We then present another algorithm called PA-SPTF that considers the aging factor so that starvation resistance is improved. Simulation studies show that PA-SPTF improves the performance of MEMS storage by up to 39.2% in terms of the average response time and 62.4% in terms of starvation resistance compared to the widely acknowledged SPTF algorithm. We also show that there exists a spectrum of scheduling algorithms that subsumes both the P-SPTF and PA-SPTF algorithms.

Personalized recommendation schemes for DTV channel selectors

The emergence of new TV media such as digital cable and satellite has diversified TV contents. This makes the task of finding ones favorite contents among hundreds of TV channels a time consuming job. To relieve this channel selection overhead, this paper presents personalized recommendation schemes for DTV channel selectors. The proposed recommendation schemes analyze each persons previous watching behavior in terms of recency and frequency, and then utilize this information in the control of the top-down channel selector. Simulation studies show that the proposed schemes reduce the seek cost of the DTV channel selector by up to 62.8% for the trace set we considered

Data allocation in MEMS-based mobile storage devices

MEMS-based storage is being developed as a new storage media. Due to its attractive features such as small size, shock resistance, and low-power consumption. MEMS-based storage is anticipated to be widely used for mobile consumer electronics. However, MEMS-based storage has vastly different physical characteristics compared to a traditional disk. First, MEMS storage has thousands of heads that can be activated simultaneously. Second, the media of MEMS storage is a square structure which is different from the platter structure of disks. Third, the size of a physical sector in MEMS storage is an order of magnitude different from that of a conventional disk. This paper presents a new data allocation scheme for MEMS storage that makes use of the aforementioned characteristics. This new scheme considers the parallelism of MEMS storage as well as the seek time of requests on the two dimensional square structure. Simulation studies show that the proposed scheme improves the performance of MEMS storage significantly by exploiting the high parallelism of MEMS storage.

Parallelism-Aware Request Scheduling for MEMS-based Storage Devices

MEMS-based storage is being developed as a new storage media. Due to its attractive features such as high-bandwidth, low-power consumption, and low cost, MEMS storage is anticipated to be used for a wide range of applications. However, MEMS storage has vastly different physical characteristics compared to a traditional disk. First, MEMS storage has thousands of heads that can be activated simultaneously. Second, the media of MEMS storage is a square structure which is different from the platter structure of disks. This paper presents a new request scheduling algorithm for MEMS storage that makes use of the aforementioned characteristics. This new algorithm considers the parallelism of MEMS storage as well as the seek time on the two dimensional square structure. We then extend this algorithm to consider the aging factor so that starvation resistance is improved. Simulation studies show that the proposed algorithms improve the performance of MEMS storage by up to 39.2% in terms of the average response time and 62.4% in terms of starvation resistance compared to the widely acknowledged SPTF (Shortest Positioning Time First) algorithm.

A smart elevator scheduler that considers dynamic changes of energy cost and user traffic

With the recent advances in energy-aware building technologies, the electricity usage of a smart building is detected every moment and might have different costs at each time slot of a day. This article presents a new elevator scheduling algorithm for a smart building that considers the dynamic changes of electricity price and passenger traffic. The goal of our algorithm is to minimize the electricity charge without increasing passengers’ waiting time. To this end, we use a control parameter to increase the number of working elevator cars when the passenger traffic is heavy or the electricity price becomes low. In contrast, when the electricity price becomes high (i.e., peak time), the system adjusts the control parameter to reduce the number of working elevator cars. This is not a simple issue as the two goals we pursue sometimes conflict. Thus, we use an optimization technique based on genetic algorithms in the design of our scheduler. To evaluate the proposed elevator scheduling system, we conduct experiments under synthetic and realistic workload conditions. The results show that the proposed elevator scheduling system significantly saves the electricity charge of the conventional elevator scheduling system. Specifically, the average reduction in the electricity charge is 68.3% without sacrificing passengers’ waiting time.

An Adaptive Location Detection scheme for energy-efficiency of smartphones

Global Positioning System (GPS) is widely used for the Location-Based Service (LBS) of smartphones. However, GPS dramatically increases the power consumption of a smartphone due to heavy computation overhead. Cell-tower Based Localization (CBL) can be an alternative solution to perform LBS in an energy-efficient way; but its adoption is limited due to the low positioning accuracy. This paper presents a new location estimation scheme for smartphones called Adaptive Location Detection (ALD). ALD adaptively detects the location of a smartphone considering the category of applications executed, movement pattern of a user, and the battery level. Specifically, ALD categorizes applications according to the required level of positioning accuracy, and then adaptively utilizes GPS and CBL. ALD also takes different actions according to the movement pattern of a user and the remaining battery level of the smartphone. To assess the effectiveness of the proposed scheme, we perform simulations under five location based applications and six scenarios. The evaluation results show that ALD reduces the energy consumption of GPS by 49.5% on average. Nevertheless, it satisfies the accuracy requirement of each situation.

Popularity and adjacency based prefetching for efficient IPTV channel navigation

As Internet Protocol Television (IPTV) is becoming increasingly popular, users can enjoy a variety of contents from hundreds of IPTV channels. However, the increased number of channels makes users even more difficult to find their desired contents. Along with this problem, the channel zapping time of IPTV incurs serious user-perceived latency. To alleviate these problems, this paper presents a new channel prefetching scheme that can be implemented on the IPTV set-top box. The scheme basically fetches the contents of adjacent channels before they are actually requested, similar to the adjacency-based prefetching scheme. However, it differentiates the number of channels to be prefetched among up-direction and down-direction channels based on the popularity of channels. Trace-driven simulations with various experimental conditions show that the proposed prefetching scheme performs better than existing prefetching schemes in terms of the prefetching hit rate to find the desired channel. Specifically, when the number of IPTV channels is 150, the average improvement of the hit rate is 22.3% compared to the adjacency-based prefetching scheme.

An Energy-Efficient Positioning Scheme for Location-Based Services in a Smartphone

As Location-Based Services (LBSs) are widely used in smartphone applications, Global Positioning System (GPS) becomes one of the main sources of a smartphones energy consumption. This paper presents an energy-efficient positioning scheme for smartphones called EEPS (Energy-Efficient Positioning Scheme). EEPS adaptively performs the positioning of a smartphone considering the interesting regions of a user, accuracy requirement of each application executed within the smartphone, and the battery level. Simulations with various real applications and scenarios show that EEPS reduces 50.1% of energy consumption compared to GPS. Nevertheless, it satisfies the accuracy requirement of each application.