Cloud, Fog, or Mist in IoT? That Is the Question

Abstract

Internet of Things (IoT) has been commercially explored as Platforms as a Services (PaaS). The standard solution for this kind of service is to combine the Cloud computing infrastructure with IoT software, services, and protocols also known as CoT (Cloud of Things). However, the use of CoT in latency-sensitive applications has been shown to be unfeasible due to the inherent latency of cloud computing services. One proposal to solve this problem is the use of the computational resources available at the edge of the network, which is called Fog computing. Fog computing solves the problem of latency but adds complexity to the use of these resources due to the dynamism and heterogeneity of the IoT. An even more accentuated form of fog computing is Mist computing, where the use of the computational resources is limited to the close neighborhood of the client device. The decision of what computing infrastructure (Fog, Mist, or Cloud computing) is the best to provide computational resources is not always simple, especially in cases where latency requirements should be met by CoT. This work proposes an algorithm for selecting the best physical infrastructure to use the computational resource (Fog, Mist, or Cloud computing) based on cost, bandwidth, and latency criteria defined by the client device, resource availability, and topology of the network. The article also introduces the concept of feasible Fog that limits the growth of device search time in the neighborhood of the client device. Simulation results suggest the algorithm’s choice adequately attends the client’s device requirements and that the proposed method can be used in IoT environment located on the edge of the network.

A comparative node evaluation model for highly heterogeneous massive‐scale Internet of Things‐Mist networks

“…Although there is a paucity of literature on the problem of resource discovery in highly heterogeneous Mist networks, the available articles are briefly reviewed. Vasconcelos et al 10 introduce a model to use the computational resources in three different layers of the Cloud, Fog, and Mist. They evaluate the cost, bandwidth and latency of each layer based on available resources and topology of the network to determine the best layer to execute tasks.…”

“…Because there are several layers available, its not clear which layer is best utilized to achieve these goals in different situations. 10 When comparing the Edge computing paradigms, the research to date has focused on layers in the network infrastructure devices (Edge and Fog) to execute tasks, rather than resource-rich end-user devices. Although network infrastructure devices are appropriate to execute tasks and host services, there are reasons to encourage the pushing of execution closer to the end-users:…”

“…19 Although providing resources on the edge of the network is a solution, there is no logical reason to ignore the free resources of IoT nodes that are more accessible, free, and flexible. 10 Despite the efficient promise of Mist computing, it still suffers from several significant drawbacks. There is not yet much research on implementing Mist computing and a lack of standards for how to do it.…”

“…While each of these paradigms is concerned with leveraging the computational resources close to end‐users in order to perform computation tasks, they achieve this in different layers of the network. Because there are several layers available, its not clear which layer is best utilized to achieve these goals in different situations 10 …”

“…Evidence shows that the future of IoT networks belongs to the network paradigms that can use and manage the available network resources efficiently in all layers, because of the volume of data that needs to be processed 19 . Although providing resources on the edge of the network is a solution, there is no logical reason to ignore the free resources of IoT nodes that are more accessible, free, and flexible 10 …”