An Aggregatable Name-Based Routing for Energy-Efficient Data Sharing in Big Data Era

Abstract

Big data strongly demands a network infrastructure having the capability to efficiently collect, process, cache, share, and deliver the data, instead of simple transmissions. Such network designs show the requirements of energy efficiency, availability, high performance, and data-aware intelligence. To meet these requirements, we adopt the information-centric networking (ICN) approach, where data are retrieved through names and in-network caching is utilized. However, as the typical existing ICN architectures, content centric network (CCN) cannot efficiently utilize the caches for data sharing because of the on-path caching strategy, and network of information (NetInf) demonstrates the resolution latency for data retrievals. To design an efficient and effective ICN architecture for big data sharing, we combine the strong points of CCN and NetInf, where information islands (IOIs) and management plane are utilized for direct data retrieval and global data discovery, respectively. We provide a reference architecture and propose an aggregatable namebased routing (ANBR), which can naturally enable consumers to retrieve the closest copy of information. In this network, each piece of data can be cached at one IOI at most once, which greatly improves the efficiency of cache usages. The consumers first try to retrieve the data in the local IOI, and then try to globally retrieve it from the closest IOI, holding the copy of the data if necessary. We investigate the impact from the key factor, IOI size, to the energy consumption of ANBR. It shows that energy consumption first decreases and then increases as the IOI size increases, and the optimized IOI size can be found for deployment. Furthermore, we study the relation between the optimized IOI size and the average retrieval times for the data. The result shows that the optimized IOI size increases as the average retrieval times increase.INDEX TERMS Big data, information centric network, CCN, energy-efficiency. VOLUME 3, 2015 2169-3536 2015 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. 955 R. Li et al.: ANBR for Energy-Efficient Data Sharingthe Internet the difficulties in providing the highly available services for these applications [25].Therefore, it is time to re-consider the network infrastructure design for data sharing applications in the era of big data. We summarize four design requirements as follows. A. ENERGY-EFFICIENCYThe network should reduce redundant and duplicate traffic to optimize energy consumptions in data transmissions. It also should enable the data to be retrieved from the closest data copy holder. For the current implementation with data centers, the same data need to be delivered from data center, which might be far away, to a set of users one by one, which brings out large duplicate traffic overhead. B. AVAILABILITYThe network...

Big Data Redux: New Issues and Challenges Moving Forward

“…One study [24] reiterated that organizations need to develop criteria to assess the relevance and criticality of streaming data when they cannot afford to store it all. Another paper [38] proposed a model for designing and developing energyefficient networks that may alleviate some of the bottlenecks in organizations that receive big data in near real-time. One paper [39] noted that "rapid increases in the volume and velocity of data streaming pose technical challenges for capturing, storing, and analyzing massive amount of data".…”

Near-Optimal Data Structure for Approximate Range Emptiness Problem in Information-Centric Internet of Things

“…where c = u/(nL). The approximation in (18) comes from the fact that u n in the scenario of the approximate range emptiness problem, see [4]. Considering that we have log 2 (L) queries for each e i , and there are n elements, we thus have at most log 2 (L)n(ε+(1−ε) 1 c ) query errors, and each bit in array D can correct exactly one error.…”

Edge Computing for the Internet of Things: A Case Study

“…Information-centric networking (ICN) [9] is a similar approach for enhancing the Internet infrastructure to explicitly support content-based routing and forwarding. However, the CDN and ICN paradigms are limited to non-interactive content [10]; for instance, IoT data can be cached at the edge of the network [11,12]. On the other hand, edge computing servers also provide computational capabilities and can host interactive applications that support user mobility.…”