Skip to main content

Hitachi

Corporate InformationResearch & Development

January 27, 2015

Report from Presenter


Photo 1: Presentation scene

The 3rd Asian Conference on Information Systems (ACIS2014) was held in Nha Trang, Vietnam, from Dec. 1 to Dec. 3, 2014. Many researchers, professionals, and practitioners discuss various current issues on information systems in ACIS. In ACIS2014 this year, there were 80 presentations including oral and poster.

We, Hitachi, made 10 oral presentations, and this report shows the two of them, "SCC: A Method of Shared File Cache for File Clone Function for Fast Virtual Machine Boot", and "Risk-based Method for Data Redundancy Determination to Improve Replica Capacity Efficiency."


Fig. 1: Overview of a method for shared file cache
Enlarge

First, we propose a method for shared file cache for fast virtual machine boot. The method reduces the number of I/Os issued to a storage device when virtual machines (VMs) are started simultaneously. When administrators deploy the VMs, the VMs are deployed by using the file clone function to easily and fast deploy these VMs. The proposed method utilizes the structure of the cloned files to share file cache. In this way, the method can mitigate I/Os issued to disk. According to the evaluation, we prove that the method reduces dramatically I/Os when multiple VMs are started simultaneously.


Fig. 2: Overview of a method to balance
availability and replica amounts

Enlarge

Next, we also proposed a method to reduce replica data amounts for a disaster-resilient storage system, which consists of storage nodes located in many different sites. The proposed method determines the number of replicas based on the disaster risk of each site. The method estimates the disaster risk value of each site improved by replicating its data, and prevents a node from replicating its data if the node reaches an acceptance level given by an administrator. This enables only high-risk sites to create many replicas, therefore, the storage system reduces the total amount of replicas with maintaining the data safety against the assuming disasters.
This research on the disaster-resilient storage system is supported as "Research and Development on Highly-functional and Highly-available Information Storage Technology", sponsored by the Ministry of Education, Culture, Sports, Science and Technology in Japan, and we are collaborating and researching with Tohoku University and Hitachi Solutions East Japan, Ltd.

(By KAMEI Hitoshi & MATSUMOTO Shinya)

Related Papers

  • Hitoshi Kamei, Osamu Yashiro, and Takaki Nakamura, "SCC: A Method of Shared File Cache for File Clone Function for Fast Virtual Machine Boot", In Proceedings of the 3rd Asian Conference on Information Systems (ACIS 2014), pp. 545-551, 2014.
  • Shinya Matsumoto, Takaki Nakamura, and Hiroaki Muraoka, "Risk-based Method for Data Redundancy Determination to Improve Replica Capacity Efficiency", In Proceedings of the 3rd Asian Conference on Information Systems (ACIS 2014), pp. 529-536, 2014.
  • Page top