Application-aware resource provisioning in a heterogeneous Internet of Things
| Autor: | Massimo Tornatore, Eric Sturzinger, Biswanath Mukherjee |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2017 |
| Předmět: |
Computational complexity theory
IoT traffic Computer science Internet of Things Cloud computing 02 engineering and technology Solid modeling dynamic lightpath reallocation 03 medical and health sciences Bandwidth cloud computing metropolitan area networks telecommunication computing telecommunication network planning telecommunication traffic wide area networks MAN service provider WAN application-aware resource provisioning heterogeneous Internet of Things hybrid fog-cloud architecture wide area network Computational modeling Computer architecture Delays Wide area networks 0202 electrical engineering electronic engineering information engineering 030504 nursing business.industry 020206 networking & telecommunications Functional requirement Provisioning Service provider Application profile Wide area network 0305 other medical science business Computer network |
| Zdroj: | ONDM |
| Popis: | Internet of Things (IoT) traffic will become increasingly heterogeneous not only in terms of traditional metrics as required bandwidth and maximum latency, but also in terms of functional requirements such as compute power and temporary storage. Sophisticated planning and engineering approaches must be adopted by service providers to account for this heterogeneity, inherent in IoT applications. Metropolitan Area Networks (MANs) are ideally suited to manage and implement resource provisioning of heterogeneous IoT application traffic and, as a result, possess a unique ability to conserve MAN and Wide Area Network (WAN) bandwidth costs. We propose a novel comprehensive MAN resource provisioning model in a hybrid fog-cloud architecture which decouples compute and storage functions while accounting for traffic of a set of heterogeneous parameterized application profiles. This is intended to assist the MAN service provider to minimize the total operational cost of provisioning IoT traffic demands as well as provide a framework for dynamic lightpath reallocation within the MAN. The model demonstrates which application profile and topological parameters have the most significant effect on the individual cost components. As a result of the model, we demonstrate that optimal resource provisioning, i.e. whether functions are placed in the fog or cloud, depends heavily on application computational complexity, compression factor, and latency budget, as well as proportions of local and global traffic. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|