Exploring potential implementations of PCE in IoT world

Abstract

The recently coined Internet of Things (IoT) paradigm leverages a large volume of heterogeneous network elements (NEs) demanding broad connectivity anywhere, anytime and anyhow, fueling the deployment of innovative Internet services, such as Cloud or Fog Computing, Data Center Networks (DCNs), Smart Cities or Smart Transportation. The proper deployment of these novel Internet services is imposing hard connectivity constraints, such as high transmission capacity, reliable communications, as well as an efficient control scheme capable of enabling an agile coordination of actions in large heterogeneous scenarios. In recent years, novel control schemes, such as the so-called Path Computation Element (PCE) has gained momentum in the network research community turning into real PCE implementations. Indeed, there is a wealth of studies assessing the PCE performance, clearly showing the potential benefits of decoupling routing control tasks from the forwarding nodes. Nevertheless, recognized the need for a control solution in IoT scenarios, there is no much published information analyzing PCE benefits in these IoT scenarios. In this paper, we provide an insight particularly demonstrating how the PCE may gracefully provide support to the service composition in an agile manner, handling the specific constraints and requirements found in IoT scenarios. To this end, we propose a novel PCE strategy referred to as Service-Oriented PCE (SPCE), which enables network-aware service composition.

Towards a proper service placement in combined Fog-to-Cloud (F2C) architectures

“…In short, historical customer's behavior data is utilized to grant both higher priority and better quality to customers with lower probability of requesting service reallocation. An extension of the PCE [31], so-called Serviceoriented PCE (S-PCE), is presented in the position paper [36]. The presented S-PCE architecture is focused on service selection and allocation in mobile clouds and proclaims, albeit conceptually, to address important issues caused by the mobility, volatility and heterogeneity of mobile network elements -different from the work described in [30], where PCE is restricted to static clouds, as described in section 2.1.…”

“…On the other hand, although works such as [35] and [37] focus on the allocation of IT resources, the resource volatility is not considered and would be critical for many applications, such as for smart cities. Authors in [33] aims at achieving low delay on service execution, however the focus is on data acquisition delay only, whereas [36] proposes a centralized control, which on the other hand can introduce additional delay until queries reply. Regarding the importance of consideration of heterogeneous resources, [33] focuses on the heterogeneity of the data type obtained by distinct data resources, whilst [34] considers energy availability only, while assuming that all other resources are unlimited.…”

“…Other works, such as [33] and [36], present use cases for service orchestration, albeit without the corresponding algorithms. It should also be noted that none of the current work focus on virtualized resource management in fog scenarios.…”

Mechanisms for service-oriented resource allocation in IoT

“…what is the best path). To cope with these issues, we propose using an extension of PCE enriched with ILSA, hereon referred to as a Service-oriented PCE (SPCE) [10]. This section starts by presenting a preliminary approach for enhancing PCE with ILSA schemes.…”

“…The relevant is the "What" (the service to be provided) rather than the "Who" or the "Where" (the LOC of the node who will provide the service). Furthermore, from a technological perspective, the decoupling of functionalities envisioned by the serviceoriented communication paradigm enables long-term scalability, since the impact on new Internet applications is more noticeable at the service layer and less at the network layer, i.e., the service layer focuses on the composition of the service, whereas the network layer solely focuses on providing enough bandwidth to set up the demanded services and network features, such as mobility and Traffic Engineering (TE) [10].…”