دانلود رایگان مقاله جایگذاری داده های لیتنسی آگاه سبزمراکز داده

Abstract

Large-scale Internet applications provide service to end users by routing service requests to geographically distributed data centers. Two concerns exist in service provisioning by the data centers. One is that users require to experience low latency while accessing data from data centers. The other is to reduce the energy consumed by network transport and the servers in the data centers. In this paper, we tackle the problem of green data placement in data centers to strike a tradeoff among access latency, energy consumption of data centers and network transport. We propose two request-routing algorithms, GLDP-NS (Green Latency-aware Data Placement - No consideration of the current data placement Status of the server) and GLDP-WS (Green Latency-aware Data Placement - With consideration of the current data placement Status of the server). We show that the green latency-aware data placement problem is NP-complete and algorithm GLDP-NS is a 3-approximation algorithm for the data placement problem without considering the data placement status of the server. We evaluate the performance of the proposed algorithms through simulations, and the simulation results demonstrate that the proposed algorithms can achieve good integrated cost performance of the latency, the energy consumption of data centers and network transport.

5. Conclusions

Large-scale Internet applications provide service to end users by routing service requests to geographically distributed data centers. Currently, the data centers and the network transport that power the applications consume significant worldwide electricity supply. At the same time, latency is an important concern for the end users. In this paper, we tackled the problem of green latency-aware data placement in the data centers. The objective was to reduce the energy consumed by network transport and the servers in the data centers, while reducing the access latency. We proposed two request-routing algorithms, GLDP-WS (Green Latency-aware Data Placement - With consideration of current data placement Status of the server) and GLDP-NS (Green Latency-aware Data Placement - No consideration of current data placement Status of the server), to strike a tradeoff among the three factors above during data placement. GLDP-NS is a 3-approximation algorithm for the data placement problem without consideration of the server data placement status. We also conducted experiments through simulations. Experimental results demonstrate that the proposed algorithms can achieve good integrated cost performance of the latency, the energy consumption of data centers and network transport.