China
Africa
Explore chapters and articles related to this topic
Storage-Centric Networking
Published in M. Bala Krishna, User-Centric and Information-Centric Networking and Services, 2019
M. Bala Krishna, Daniel Corujo
Software-defined storage systems (SDS) [19] reduce storage and retrieval complexities in cloud computing applications. Storage virtualization in SDS aims to achieve (i) automation of shared storage systems with efficient data accessing policies and (ii) resource optimization by using object storage systems to sustain the large data growth. An object storage system with proxy and storage nodes allows the concurrent data access of multiple tenants with varying data filters. The advantage of data filters is that they allow the tenants to use only a fraction of a large data set to complete the task. This approach considerably increases the data flow and reduces the download traffic of tenants. The multi-stage and ready-to-use dynamic data accessing policies allow the flexibility to read or write the user codes based on object requests. The data policies in terms of domain-specific language are given as follows: State the options for DO and WHEN clauses of data caching.Set the data compression for large data files exceeding the limits.Modify the bandwidths for current I/O operations.
iSDS: a self-configurable software-defined storage system for enterprise
Published in Enterprise Information Systems, 2018
Wen-Shyen Eric Chen, Chun-Fang Huang, Ming-Jen Huang
Software-defined storage (SDS) is another representative advance in storage. SDS is a way of designing and constructing storage by creating a software stack above commodity hardware (Chen, Huang, and Huang 2014; Thereska et al. 2013). This software stack implements storage features that are traditionally provided by custom-made field-programmable gate array on traditional embedded storage systems. Therefore, the cost of constructing SDS is much lower than the traditional embedded storage system.