57 research outputs found
Intra- and inter-server smart task scheduling for profit and energy optimization of HPC data centers
No full textServers in a data center are underutilized due to over-provisioning, which contributes heavily toward the high-power consumption of the data centers. Recent research in optimizing the energy consumption of High Performance Computing (HPC) data centers mostly focuses on consolidation of Virtual Machines (VMs) and using dynamic voltage and frequency scaling (DVFS). These approaches are inherently hardware-based, are frequently unique to individual systems, and often use simulation due to lack of access to HPC data centers. Other approaches require profiling information on the jobs in the HPC system to be available before run-time. In this paper, we propose a reinforcement learning based approach, which jointly optimizes profit and energy in the allocation of jobs to available resources, without the need for such prior information. The approach is implemented in a software scheduler used to allocate real applications from the Princeton Application Repository for Shared-Memory Computers (PARSEC) benchmark suite to a number of hardware nodes realized with Odroid-XU3 boards. Experiments show that the proposed approach increases the profit earned by 40% while simultaneously reducing energy consumption by 20% when compared to a heuristic-based approach. We also present a network-aware server consolidation algorithm called Bandwidth-Constrained Consolidation (BCC), for HPC data centers which can address the under-utilization problem of the servers. Our experiments show that the BCC consolidation technique can reduce the power consumption of a data center by up-to 37%
Exploring Wireless Data Center Networks: Can They Reduce Energy Consumption While Providing Secure Connections?
No full textData centers have become the digital backbone of the modern world. To support the growing demands on bandwidth, Data Centers consume an increasing amount of power. A significant portion of that power is consumed by information technology (IT) equipment, including servers and networking components. Additionally, the complex cabling in traditional data centers poses design and maintenance challenges and increases the energy cost of the cooling infrastructure by obstructing the flow of chilled air. Hence, to reduce the power consumption of the data centers, we proposed a wireless server-to-server data center network architecture using millimeter-wave links to eliminate the need for power-hungry switching fabric of traditional fat-tree-based data center networks. The server-to-server wireless data center network (S2S-WiDCN) architecture requires Line-of-Sight (LoS) between servers to establish direct communication links. However, in the presence of interference from internal or external sources, or an obstruction, such as an IT technician, the LoS may be blocked. To address this issue, we also propose a novel obstruction-aware adaptive routing algorithm for S2S-WiDCN.
S2S-WiDCN can reduce the power consumption of the data center network portion while not affecting the power consumption of the servers in the data center, which contributes significantly towards the total power consumption of the data center. Moreover, servers in data centers are almost always underutilized due to over-provisioning, which contributes heavily toward the high-power consumption of the data centers. To address the high power consumption of the servers, we proposed a network-aware bandwidth-constrained server consolidation algorithm called Network-Aware Server Consolidation (NASCon) for wireless data centers that can reduce the power consumption up to 37% while improving the network performance. However, due to the arrival of new tasks and the completion of existing tasks, the consolidated utilization profile of servers change, which may have an adverse effect on overall power consumption over time. To overcome this, NASCon algorithm needs to be executed periodically. We have proposed a mathematical model to estimate the optimal inter-consolidation time, which can be used by the data center resource management unit for scheduling NASCon consolidation operation in real-time and leverage the benefits of server consolidation.
However, in any data center environment ensuring security is one of the highest design priorities. Hence, for S2S-WiDCN to become a practical and viable solution for data center network design, the security of the network has to be ensured. S2S-WiDCN data center can be vulnerable to a variety of different attacks as it uses wireless links over an unguided channel for communication. As being a wireless system, the network has to be secured against common threats associated with any wireless networks such as eavesdropping attack, denial of services attack, and jamming attack. In parallel, other security threats such as the attack on the control plane, side-channel attack through traffic analysis are also possible. We have done an extensive study to elaborate the scope of these attacks as well as explore probable solutions against these issues. We also proposed viable solutions for the attack against eavesdropping, denial of services, jamming, and control-plane attack. To address the traffic analysis attack, we proposed a simulated annealing-based random routing mechanism which can be adopted instead of default routing in the wireless data center
An Energy-Efficient, Wireless Top-of-Rack to Top-of-Rack Datacenter Network Using 60GHz Links
No full textIncreasing interposer utilization: A scalable, energy efficient and high bandwidth multicore-multichip integration solution
No full textModeling VHF air-to-ground multipath propagation channel and analyzing channel characteristics and BER performance
No full textPerformance Evaluation of a Power-Efficient and Robust 60 GHz Wireless Server-to-Server Datacenter Network
No full text
Reducing Power Consumption of Datacenter Networks with 60GHz Wireless Server-to-Server Links
No full text- …
