Current Studies

1. Cloud Storage

1.1 Network coding for inter-cloud storage

Background

Inter-cloud (or Cloud of Clouds) is viewed as the next revolution in the cloud computing paradigm wherein the computational and data infrastructure for handling scientific, business and enterprise applications spans across multi-type and multi-brand clouds and inter-cloud storage is exactly one of key parts of inter-cloud. Due to high transmission cost among inter-cloud storage nodes, it makes the difficulty of promotion for inter-cloud storage.
To reduce the bandwidth cost consuming by system operations of inter-cloud storage, we start our studies based on inter-cloud storage, adopt network coding as the main technique, and implement a network-coding-based system architecture using cloud computing resources. Based on it, we design optimal repair schemes, scaling algorithms and efficient data fault tolerance transition mechanisms. Our goal is to reduce the bandwidth cost consumed by data repair, data scaling and fault tolerance transition of inter-cloud storage so as to lower running costs of inter-cloud storage, and promote the application and diffusion of inter-cloud storage.
As shown in Figure 1, inter-cloud storage system includes three main modules: client module, proxy module and cloud storage module. The proxy module is responsible for the encoding/decoding functions and managing metadata. As shown in Figure 2, compared with Reed-Solomon codes, network coding techniques can significantly reduce the bandwidth cost consumed by repair operations of inter-cloud storage.

Figure 1. Network coding based inter-cloud storage

Reed-Solomon codes (the above sub-figure)

Network codes (the below sub-figure)

Figure 2. Reed-Solomon codes vs. Network codes

Papers

TC '14, Infocom '13, Fast '12 (See Publications for details).

1.2 Replica placement for inter-cloud storage

Work in progress.

2. Heterogeneous/Hierarchical Storage

2.1 Regenerating coding against hierarchical failures

Work in progress.

3. RAID for In-memory NoSQL Databases

Work in progress.

4. Lineage for Big-data Storage

Work in progress.