Increasing energy consumption in server consolidation environments leads to high maintenance costs for data centers. Main memory, no less than processor, is a major energy consumer in this environment. This paper proposes a technique for reducing memory energy consumption using virtual machine scheduling in multicore systems. We devise several heuristic scheduling algorithms by using a memory power simulator, which we designed and implemented. We also implement the biggest cover set first (BCSF) scheduling algorithm in the working server system. Through extensive simulation and implementation experiments, we observe the effectiveness of the memory-aware virtual machine scheduling in saving memory energy. In addition, we find out that power-aware memory management is essential to reduce the memory energy consumption.

In a consolidated server system using virtualization, physical device accesses from guest virtual machines (VMs) need to be coordinated. In this environment, a separate driver VM is usually assigned to this task to enhance reliability and to reuse existing device drivers. This driver VM needs to be highly reliable, since it handles all the I/O requests. This paper describes a mechanism to detect and recover the driver VM from faults to enhance the reliability of the whole system. The proposed mechanism is transparent in that guest VMs cannot recognize the fault and the driver VM can recover and continue its I/O operations. Our mechanism provides a progress monitoring-based fault detection that is isolated from fault contamination with low monitoring overhead. When a fault occurs, the system recovers by switching the faulted driver VM to another one. The recovery is performed without service disconnection or data loss and with negligible delay by fully exploiting the I/O structure of the virtualized system.

Although WSRF (Web Services Resource Framework) and Java-based hosting environment have been successful in dealing with the heterogeneity of resources and the diversity of applications, the current Grid middleware has several limitations to support on-demand resource provisioning effectively.

Recently grid computing has started to leverage Web services technology by proposing OGSI-standard. OGSI standard defines the grid service, which presents unified interfaces to every participant of grid. In current Glubus Toolkit3(GT3),which is an implementation of OGSI, grid service factories should be deployed manually into resources to provide grid services. However, it is necessary to dynamically allocate proper amount of resource, since the demand for resource of service provider changes over time. In this paper, we propose a architecture to enable on-demand resource provisioning. We develop universal factory service (UFS) that provides a dynamic grid service deployment mechanism and a resource broker called door service. Through the experiments, we show that grid services can adaptively exploit resources according to the request rates.

Since user-level communication architecture (ULC) provides only primitive operations for application programmers, many high-level communication layers have been developed on top of ULC. One of such high-level communication layers is the sockets interface and it can be supported over ULC architectures in several ways. The primary objective of this paper is to identify design issues and tradeoffs among these different approaches, and to quantitatively analyze their performance to understand the various costs associated with the communication. In this paper, we design and implement KSOVIA, a kernel-level sockets layer over VIA, and compare it with the existing approaches such as a user-level sockets layer over VIA and an IP emulation layer over VIA. Our measurement results show that using an IP emulation layer exhibits the worst performance in terms of latency and bandwidth and a user-level sockets layer is useful for latency-sensitive applications. KSOVIA is found to be effective for applications, which require high bandwidth or the full compatibility with the sockets interface.