Loading...

An Efficient and Load Balanced I/O Cache Architecture for Hyper-Converged Infrastructures

Ahmadi, Sina | 2022

133 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55392 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Asadi, Hossein
  7. Abstract:
  8. In new generations of data storage systems, there is a tendency to integrate processing and storage elements. This is currently recognized with the commercial name of Hyperconverged Infrastructure, or HCI for short. The goal of this infrastructure is to meet the increasing requirements of data-intensive applications. As a new paradigm, HCI needs to incorporate different performance improvement techniques such as I/O caching. Contemporary I/O caching schemes are optimized for traditional storage architectures and do not prove optimal in a Hyperconverged Infrastructure. The main challenge of I/O cache in this kind of infrastructure is the distribution of its resources over an array of nodes (servers) which leads to two key problems: 1) imbalanced I/O cache requirements such that some nodes host virtual machines which heavily depend on cache to provide good performance, while others mostly have idle cache space available, and 2) imbalanced I/O traffic and load in a way that some nodes experience burst of requests which leads to long queue time and performance drop, while other nodes are not fully utilizing their I/O cache capacity. This imbalance has severe impact on performance and latency of the I/O cache. In this research, we propose an efficient and load-balanced I/O cache architecture, specifically designed for Hyperconverged Infrastructure. Our proposed architecture manages the SSD cache resources in such a way to improve performance. The proposed architecture: 1) dynamically reconfigures cache resources and distributes them throughout the array, and 2) by dynamic reallocation of I/O cache to virtual machines, it tries to minimize network traffic and balance SSD load. This architecture is implemented on real-world servers and tested with real workload traces. After implementation, it is shown that our proposed architecture leads to 2.8 times and 23% improvement in average and worst-case latencies, respectively. Furthermore, cache space utilization increases by 76% in our experiments.
  9. Keywords:
  10. Load Balancing ; Input/Output Cache ; Data Storage ; Service Quality ; Hyperconverged Infrastructure

 Digital Object List

 Bookmark