An efficient cloud storage system for tele-health services

Healthcare service is a critical aspect of our daily lives. Enabled by technologies such as wearable devices and wireless sensor networks, tele-health has becoming a promising new field in IT industry. Wearable devices, which detect real-time human body conditions, form body sensor networks (BSNs) for patients. In a cloud-enabled tele-health ecosystem, health data are collected by the BSN and sent to mobile devices such as smart phones and tablets. These embedded devices process the data and forward them to remote data centers. Due to the energy and time constraints of embedded systems, the effectiveness of storage systems become a critical issue. For years, memory technologies such as SRAMs and DRAMs have been widely used in computer systems. SRAMs are fast while DRAMs have high density. However, SRAMs have the disadvantage of power leakage and low density. DRAMs are slower in read and write operations. New memory technology for embedded tele-health is needed. In the paper, we propose a hybrid memory system for embedded tele-health. We combine phase-change memory PCM with flash memory to meet energy and latency requirement while reducing capital expenditure. Moreover, the data allocation and storage on server side is also a challenging problem in tele-health. Effective storage system designs are desired to efficiently store and manage health care data from users. Therefore, in the paper, we design a ecosystem for tele-health including the memory storage for embedded devices and data storage for tele-health data centers. To fully utilize the proposed ecosystem, we design several resource allocation algorithms with dynamic programming and heuristics. The experiments show that our approaches can achieve up to 30% performance enhancement compared to greedy approaches.

[1]  Norman P. Jouppi,et al.  CACTI: an enhanced cache access and cycle time model , 1996, IEEE J. Solid State Circuits.

[2]  Enhong Chen,et al.  Task Allocation on Nonvolatile-Memory-Based Hybrid Main Memory , 2013, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[3]  Trevor Mudge,et al.  MiBench: A free, commercially representative embedded benchmark suite , 2001 .

[4]  Meikang Qiu,et al.  Variable Partitioning and Scheduling for MPSoC with Virtually Shared Scratch Pad Memory , 2010, J. Signal Process. Syst..

[5]  M. Breitwisch.  Phase Change Memory , 2008, 2008 International Interconnect Technology Conference.

[6]  S. Lai,et al.  Current status of the phase change memory and its future , 2003, IEEE International Electron Devices Meeting 2003.

[7]  Meikang Qiu,et al.  Resource allocation robustness in multi-core embedded systems with inaccurate information , 2011, J. Syst. Archit..

[8]  C. Mazure,et al.  Ultra-scaled Z-RAM cell , 2008, 2008 IEEE International SOI Conference.

[9]  Zhi Chen,et al.  Energy-Aware Data Allocation With Hybrid Memory for Mobile Cloud Systems , 2017, IEEE Systems Journal.

[10]  Ricardo Bianchini,et al.  Page placement in hybrid memory systems , 2011, ICS '11.

[11]  Wei-Che Tseng,et al.  Optimizing Data Allocation and Memory Configuration for Non-Volatile Memory Based Hybrid SPM on Embedded CMPs , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[12]  Keke Gai,et al.  Phase-Change Memory Optimization for Green Cloud with Genetic Algorithm , 2015, IEEE Transactions on Computers.

[13]  Onur Mutlu,et al.  Enabling Efficient and Scalable Hybrid Memories Using Fine-Granularity DRAM Cache Management , 2012, IEEE Computer Architecture Letters.