On Maximizing Task Throughput in IoT-Enabled 5G Networks Under Latency and Bandwidth Constraints

Abstract

Fog computing in 5G networks has played a significant role in increasing the number of users in a given network. However, Internet-of-Things (IoT) has driven system designers towards designing heterogeneous networks to support diverse demands (tasks with different priority values) with different latency and data rate constraints. In this paper, our goal is to maximize the total number of tasks served by a heterogeneous network, labeled task throughput, in the presence of data rate and latency constraints and device preferences regarding computational needs. Since our original problem is intractable, we propose an efficient solution based on graph-coloring techniques. We demonstrate the effectiveness of our proposed algorithm using numerical results, real-world experiments on a laboratory testbed and comparing with the state-of-the-art algorithm.

FPGA-Based Hardware/Software Co-Design of a Bio-Inspired SAT Solver

“…As well as the results in Table 2, the linear increase in Slices# and the clock degradation are seen along with the instance complexity (i.e., the variables# and clauses#). The execution time is 0.027-0.16ms, which indicate that our work can handle, for example, the 5G network scheduling problem composed of 50 to 100 nodes (covering 250-500 IoT devices [16]) considering that each node (i.e., fog access point) has to response to its respective IoT devices within an interval of 1ms. These results satisfy the requirement of the target application and demonstrate the practicality of our work.…”

“…In the first evaluation, in order to demonstrate the effectiveness of our work over state-of-the-arts, we used six randomly generated instances composed of 100 to 250 variables (specified with a name ''uf<variables#>-<index#>'') including ones that were also used in [17], [21]. Then, in the second evaluation, in order to show the efficiency of our work in handling real-life applications, we used three SAT-encoded flat graph colouring instances composed of 150 to 300 variables (specified with a name ''flat<vertices#>-<index#>''; the vertices# represent the complexity of the graph colouring problem before encoding to SAT) since some real-life applications can be expressed in a graph colouring problem [16]. In both evaluations, we discuss the results of our work in terms of the performance (i.e., iterations#, clock frequency, and execution time), the resource usage (i.e., Slices# that represent the circuit area), and the area-delay-product (ADP) 4 which is calculated by the product of Slices# and the execu-tion time.…”

“…The second evaluation was conducted to show the efficiency of our work in handling real-life applications such as 5G network scheduling problems [16]. Because those problems can be potentially formulated as group colouring problems (GCPs), 5 we selected three instances that represent GCPs (flat50-1, flat75-10, and flat100-1).…”

“…Hence, for now, they are more well-suited to cloud settings rather than for edge devices. Contrary, although most SAT solvers handle Boolean variables, generalization/encoding techniques to handle combinatorial problems have been well studied [16]. Moreover, stochastic SAT solvers are lightweight, yet they can achieve searching performance comparable to their heavier stochastic alternatives.…”

Survey of Radio Resource Management in 5G Heterogeneous Networks

“…Fulfillment of these user requirements must provide seamless services and real-time response; hence, high latency cannot be tolerated. It is crucial to maintain the induced latency at a very minimum level (less than 1 ms) [14,87,88] and thus resource management techniques must address the implementational and computational complexities at the best possible level. The complexities of radio resource management are summarized in Table V.…”