Norifumi Nishikawa

Memory-based architecture for distributed WWW caching proxy

Abstract In networks with heavy WWW traffic, the disk I O performance can become a bottleneck for caching proxies. To solve this problem, we propose a memory-based architecture for WWW caching proxies. The features of the proposed architecture are (1) a cache contents control that retrieves only frequently accessed WWW pages, and (2) an automatic distribution mechanism that enables efficient cache space sharing among multiple caching proxies. A statistical analysis shows that our approach reduces the required cache space. It consumes only 1 10 the cache space of current typical proxies while retaining the same cache hit rate. The reduced cache space enables the use of fast DRAMs in the caching proxies, and solves the disk I O problem.

A proposal of storage power control method with data placement in an environment using many HDDs

In recent years, the scale of datacenters has become larger due to the explosive increase in the amount of digital data. As a result, the growth of energy consumption is an important factor in the management use cost of datacenters. Storing and processing such large volumes of data by database applications are the core technologies in this Big Data era. However, storage accounts for a significant percentage of a datacenters energy consumption. Therefore, we try to reduce the energy of storage to save on the total cost of datacenters. The purpose of this study is to reduce the energy consumption of storage while minimizing the deterioration of application performance. Although many methods for storage energy saving have been discussed, since it is difficult to control it efficiently only at the storage level, we have investigated the storage power control mechanism on middleware (database) layer. In this paper, we use TPC-H (a database benchmark) as an application example of data processing. We evaluate a data placement control method of storage proposed for energy saving in the database run-time processing suitable for a large-scale environment with many HDDs.

Energy aware RAID configuration for large storage systems

Power consumption of storage at data centers is increasing rapidly. Large storage facilities have various RAID configurations incorporating different RAID levels, numbers of drives, and media types. Nevertheless, few discussions of RAID configurations have been pursued from an energy saving perspective. We first investigate how different RAID configurations affect not only application performance but also power consumption of storage installations. We then present simulation results of power consumption and application performance conducted with various RAID configurations. Results show that a RAID configuration strongly impacts energy conservation based on application I/O features.

Low power management of OLTP applications considering disk drive power saving function

A power-saving management for OLTP applications has become an important task for user budgets and datacenter operations. This paper presents a novel power-saving method for multiple disk drives based on knowledge of OLTP application behaviors. We report detail analysis of power consumption of disk drives and I/O characteristics of OLTP application. We then show experimental and simulation results of our power-saving methods. Our method provides substantially lower power consumption of disk drives compared to that of a conventional OLTP environment.

Application sensitive energy management framework for storage systems

Rapidly escalating energy and cooling costs, especially those related to the energy consumption of storage systems, have become a concern for data centers, primarily because the amount of digital data that needs storage is increasing daily. In response, a multitude of energy saving approaches that take into account storage-device-level input/output (I/O) behaviors have been proposed. The trouble is that numerous critical applications such as database systems or web commerce applications are in constant operation at data centers, and the conventional approaches that only utilize storage-device-level I/O behaviors do not produce sufficient energy savings. It may be possible to dramatically reduce storage-related energy consumption without degrading application performance levels by utilizing application-level I/O behaviors. However, such behaviors differ from one application to another, and it would be too expensive to tailor methods to individual applications. As a way of solving this problem, we propose a universal storage energy management framework for runtime storage energy savings that can be applied to any type of application. The results of evaluations show that the use of this framework results in substantive energy savings compared with the traditional approaches that are used while applications are running.

Proposal of analytics software architecture with data preparation layer for fast event identification in wide-area situational awareness

The number of PMUs (Phasor Measurement Units) has increased drastically in several countries. The collected data such as voltage, current, and frequency become over hundreds terabytes in a few years. It takes long time to retrieve specific similar patterns to a query for analysis. This paper defines data preparation layer in an analytical system. That enables enormous accumulation. The layer contains a function that creates index creation based on time-series patterns and another function that quickly retrieves user required patterns. Both functions process clusters as well as string. In the experiment, the time to retrieve similar patterns to a query is measured, and four architectures of time-series data process are compared; (1) conventional architecture with no data preparation layer (2) architecture with String (3) architecture with Cluster, and (4) architecture with String and Cluster. The experiment result shows that the best one with String and Cluster is 67.7 times faster than the conventional no-data preparation layer architecture. In conclusion, the proposed data preparation layer is effective to retrieve similar time-series data patterns to the query. It is also clarified that the layer is critical to analyze historical data for predicting wide-area power disturbance.

Energy aware RAID configuration for data intensive applications in enterprise storages

Energy consumption of storage at data centres is increasing rapidly. Large storage facilities have various RAID configurations incorporating different RAID levels, numbers of drives, and media types. Nevertheless, few discussions of RAID configurations have been pursued from an energy-savings perspective. We first investigate how different RAID configurations affect not only application performance but also power consumption of storage installations. We then present the simulation results for power consumption and application performance obtained using various RAID configurations. These results reveal that a RAID configuration strongly impacts energy conservation based on application I/O features.

Application Sensitive Energy Management Framework for Storage Systems

Rapidly escalating energy and cooling costs, especially those related to the energy consumption of storage systems, have become a concern for data centers, primarily because the amount of digital data that needs storage is increasing daily. In response, a multitude of energy saving approaches that take into account storage-device-level input/output (I/O) behaviors have been proposed. The trouble is that numerous critical applications such as database systems or web commerce applications are in constant operation at data centers, and the conventional approaches that only utilize storage-device-level I/O behaviors do not produce sufficient energy savings. It may be possible to dramatically reduce storage-related energy consumption without degrading application performance levels by utilizing application-level I/O behaviors. However, such behaviors differ from one application to another, and it would be too expensive to tailor methods to individual applications. As a way of solving this problem, we propose a universal storage energy management framework for runtime storage energy savings that can be applied to any type of application. The results of evaluations show that the use of this framework results in substantive energy savings compared with the traditional approaches that are used while applications are running.

Cache effect for power savings of large storage systems with OLTP applications

The power consumption of modern datacenters is increasing rapidly. Storage in datacenter consumes much power. Today, databases, especially those for OLTP, have become a major storage application in datacenters. Therefore, power-saving management for OLTP applications has become an important task for user budgets and datacenter operations. This paper presents a description of a novel power-saving method for large storage systems based on application I/O behavior of OLTP applications. Features of our approach are (i) measurement of actual RAID storage power consumption, (ii) analysis of I/O behavior characteristics of OLTP applications, and (iii) delayed write operation at a storage cache level based on the I/O behavior of OLTP applications. We present a measured result of power consumption of storages during an OLTP application is running, and simulations results of our power-saving methods with varying cache size of storage, which demonstrate that our method provides substantially lower storage power consumption than that of a conventional OLTP environment.

Interactive Consultation System with Asymmetrical Communication between People in Different Electronic Communities

Recently, in electronic community, information retrieval and communication have bcome popular among people with different cultures, variety of circumstances and organization in electronic communities. However, it is sometimes difficult to find suitable information and services in a short enough time for real use, because quality of information on internet is not unified and quantity of information is enormous on internet. In this paper, we discuss communication issues in electronic communities, especially asymmetrical communication between experienced experts and residents with less information. These are two view points: 1) Retrieving Heterogeneous Information Sources, and 2) Integrating Heterogeneous Information Sources.