Figure 2 - uploaded by Aluizio Rocha Neto
Content may be subject to copyright.
Source publication
Emerging Internet of Things (IoT) applications based on distributed sensors and machine
intelligence, especially in smart cities, present many challenges for network and processing infrastructure. For example, a single system with a few dozen monitoring cameras is sufficient to saturate the city’s backbone. Such a system generates massive data stre...
Contexts in source publication
Context 1
... this way, relatively powerful edge devices can perform some advanced data analytics and processing, thus reducing system response time and network traffic. Figure 2 presents the differences between the traditional organization of ...
Context 2
... created a specialization of the Data Communication Manager com-ponent as a broker to control the load balancing. Figure 20 shows the workflow nodes organization. MLO nodes are the IIM message producers, FLO nodes are the workers and one DLO is the final consumer of the processing flow. ...
Context 3
... ran simulations to process 100, 300, 600 and 1000 IIM . Figure 23 shows the results for Experiment E2. Using MELINDA, the processing time by the set of FLO nodes with capacities 40, 20 and 10 FPS was about 15% lower when compared to a RR technique. ...
Context 4
... MELINDA, the processing time by the set of FLO nodes with capacities 40, 20 and 10 FPS was about 15% lower when compared to a RR technique. However, when there is a significant disparity between the capabilities of the FLO nodes, MELINDA's time to process the messages is about 33% faster than a RR solution, as illustrated in Figure 24b. MELINDA took 1100 ms to process 100 messages while the RR technique took 1640 ms. ...
Context 5
... is the standard deviation, i.e., In each experiment, the node allocation algorithm should choose a set of nodes that could process a workflow with a demand of 200 FPS. Figure 25 presents the operational cost (equation (6.15)) calculated for each algorithm and experiment. We named our algorithm MELINDA-RA. ...
Context 6
... 13 presents the data for twenty edge nodes and twenty cloud nodes used in these tests, in which the processing capacity demand is also 200 FPS. Figure 26 shows the maximum latency from the input data source to the processing node in the cluster. Solutions with only edge nodes are the ones that obtain the lowest latency since the processors are close to the data sources -the sensors. ...
Context 7
... with only edge nodes are the ones that obtain the lowest latency since the processors are close to the data sources -the sensors. Figure 27 shows Figure 28 presents the availability of each cluster of nodes. A more significant number of edge nodes outperforms the better availability of fewer cloud nodes. ...
Context 8
... with only edge nodes are the ones that obtain the lowest latency since the processors are close to the data sources -the sensors. Figure 27 shows Figure 28 presents the availability of each cluster of nodes. A more significant number of edge nodes outperforms the better availability of fewer cloud nodes. ...
Citations
... Analysis of recent research and publications. Despite the relevance of the study of edge computing architectures, there is currently no systematization in this area, no comprehensive analysis and comparative evaluation of different architectures, although many authors, both domestic and foreign, have partially studied this issue [1][2][3][4][5][6][7][8][9][10][11]. Given the rapid development of this area and its potential for the introduction of new technologies, the availability of such a study is a scientific and practical need. ...
... To solve these problems, Neto A.R. proposed a three-tier Fog Computing architecture (Fig. 4) [7]. According to Figure 4, the basic principle of this architecture is to reduce the amount of data by processing the video stream in multiple stages. ...
The article analyzes modern architectures of edge and fog computing systems, including OpenFog, F2c2C (Cloudlet), MELINDA, and architectures based on SDN and NFV. Particular attention is given to the study of Fog Computing from the conceptual and programmatic points of view. The advantages and limitations of the studied architectures in the context of IoT application are determined. Opportunities for enhancing telecommunication systems and improving the quality of service through the use of appropriate architectures are identified. The necessity of taking into account the specific needs and features of each system when choosing the appropriate fog computing architecture is proved. The need and relevance of further development and improvement of these architectures for optimal use are substantiated.
