Yoshiko Yasuda

Concept and evaluation of X-NAS: a highly scalable NAS system

X-NAS (expandable network attached storage), a highly scalable, distributed file system designed for entry-level NAS, has been developed. It virtualizes multiple NAS systems into a single-file-system view for different kinds of clients. The core of X-NAS is a multi-protocol virtualized file system (MVFS), and its key features - a smart-code wrapper daemon, file-group mapping, and a file-handle cache - improve X-NAS scalability. X-NAS has other key features for improving the manageability on many NAS systems; namely, on-line reconfiguration, autonomous rebalancing, and automatic migration, in which files are migrated automatically and dynamically independently of file-sharing services for clients. To validate the X-NAS concept, an X-NAS prototype was designed and tested according to the NFSv2 implementation. These tests indicate that X-NAS attains a quicker response time and higher throughput than a conventional single NAS, so its cost-performance scalability is also higher.

Deadlock-free fault-tolerant routing in the multi-dimensional crossbar network and its implementation for the Hitachi SR2201

We have developed a hardware detour path selection facility for the Hitachi SR2201 parallel computer, which uses a multi-dimensional crossbar as an inter-processor network to ensure operating efficiency and high reliability when a part of the network is faulty. When this hardware facility is used, packets are transmitted to their destination along alternative paths to avoid the fault. However, changing the routing may cause deadlock. This paper describes a deadlock-free fault-tolerant routing scheme that can be used by the detour path selection facility to avoid deadlock, and its implementation for the SR2201.

Discovering configuration templates of virtualized tenant networks in multi-tenancy datacenters via graph-mining

Multi-tenant datacenter networking, with which multiple customer (tenant) networks are virtualized over a single shared physical infrastructure, is cost-effective but poses significant costs on manual configuration. Such tasks would be alleviated with configuration templates, whereas a crucial difficulty stems from creating appropriate (i.e., reusable) ones. In this work, we propose a graph-based method of mining configurations of existing tenants to extract their recurrent patterns that would be used as reusable templates for upcoming tenants. The effectiveness of the proposed method is demonstrated with actual configuration files obtained from a business datacenter network.

Discovering Cloud Operation History through Log Analysis

Management costs in private clouds will be promisingly reduced by reviewing ‘operation history,’ which is defined as a holistic view of past operation executions. Operation history provides insights into breakdown of operations: the breakdown clarifies cost-dominant operations to be improved and repetitive ones to be automated. Towards obtaining the operation history, a conventional approach relying on manual investigation is time-consuming, and another relying on agent-based monitoring is not often acceptable in sensitive mission-critical enterprise clouds. Different from these approaches, our idea is to discover the operation history by automatically analyzing ‘system logs’ that are easily accessible even in sensitive clouds. Since system logs contain only low-level debugging messages about programmatic events without direct contexts about operations, the challenge is to recover high-level operational contexts from low-level system logs. To address this challenge, we develop a method that first abstracts system logs using a pre-defined event sequence model, and then maps the abstracted events to high-level individual operations—this mapping between different contextual levels is achieved by using complementary cross-cloud reference data. Evaluation of an implementation revealed that this method reduces the time taken to discover the history by 99.9% compared to a conventional approach while achieving up to 95% correctness.

Proposal and Evaluation of Data-Model Translation for Server-Virtualization Environment to Improve Virtual-Network Configuration Management

To reduce the operational load of virtual-network operation in data centers that contain server virtualization platforms and network switches provided by multiple vendors, it would become the challenge to ease management of configurations which are different among server-virtualization environments. This paper proposes a new data-model translation method to unify data-model among multiple configurations of server-virtualization platforms. As the result of evaluation with 24 virtual machines, data-model translation overhead was less than 10% of the time to acquire virtual-machine configuration, and the network management system took 80.5 s to generate configuration information for the overall system. These results indicate that the proposed method is effective to improve the efficiency of configuration management of virtual networks.

Evidence-based context-aware log data management for integrated monitoring system

Integrated monitoring system, enabled with semi-structured datastore, is a promising solution for monitoring SaaS systems. However, according to increasing scale of SaaS systems and their long-term of service operations, the monitoring system has faced the problem in response times of log analysis and storage consumption. Our empirical observation is that the problem is primarily derived from the unselective log processing of semi-structure datastore, whereas there should be heterogeneities in log data that we can take advantage of for efficient log management. Based on this observation, we first attest this insight by investigating the usage patterns of log data in a quantitative manner with an actual dataset of log access histories obtained from a SaaS system serving to enterprise users, and we show that there are heterogeneities in required retention period of logs, response time, and amount of data, depending on log data category and its analysis scenario. Armed with the evidence found from the investigation, we design a methodology of context-aware log management, key features of which are to speculatively pre-cache the log analysis and to proactively archive ones depending on log data category and analysis scenario. Evaluation with a prototype implementation shows that the proposed method reduces the response time and the storage consumption.