Moisés R. N. Ribeiro

Keyflow: a prototype for evolving SDN toward core network fabrics

The large bulk of packets/flows in future core networks will require a highly efficient header processing in the switching elements. Simplifying lookup in core network switching elements is capital to transport data at high rates and with low latency. Flexible network hardware combined with agile network control is also an essential property for future software-defined networking. We argue that only further decoupling between the control and data planes will unlock the flexibility and agility in SDN for the design of new network solutions for core networks. This article proposes a new approach named KeyFlow to build a flexible network-fabricbased model. It replaces the table lookup in the forwarding engine by elementary operations relying on a residue number system. This provides us tools to design a stateless core network by still using OpenFlow centralized control. A proof of concept prototype is validated using the Mininet emulation environment and OpenFlow 1.0. The results indicate RTT reduction above 50 percent, especially for networks with densely populated flow tables. KeyFlow achieves above 30 percent reduction in keeping active flow state in the network.

An Analytical Approximated Solution for the Gain of Broadband Raman Amplifiers With Multiple Counter-Pumps

A novel analytical approximated solution for the gain of broadband Raman amplifiers with multiple counter-pump lasers is derived taking into account pump-pump interactions. Validation against experimental data available in the literature with two and three pumps is performed. In order to test the limits of the approximated analytical solution, signal gain is numerically obtained by simulation for Raman amplifiers built with different types of fibers and covering bandwidths up to 100 nm. The results show that our analytical solution presents a good agreement with both numerical and experimental results in most of the practical situations. This allows the reduction of the computation time in several orders of magnitude when optimizing such amplifiers.

Algorithms for FWM-aware routing and wavelength assignment

This work presents a study on the impact of the four wave mixing in dynamic optical networks. We have proposed two algorithms for routing and wavelength assignment based on the limitations imposed by four wave mixing crosstalk. When the efficiency of the effect is not very high, the algorithms present a good performance in order to minimize the four wave mixing. The results suggest that, under certain conditions, there is no loss in basing the allocation of long routes only on the amount of FWM contamination of the new path, and not on the other active paths.

Corrosion Resistant FBG-Based Quasi-Distributed Sensor for Crude Oil Tank Dynamic Temperature Profile Monitoring

This article presents a corrosion resistant, maneuverable, and intrinsically safe fiber Bragg grating (FBG)-based temperature optical sensor. Temperature monitoring is a critical activity for the oil and gas industry. It typically involves acquiring the desired parameters in a hazardous and corrosive environment. The use of polytetrafluoroethylene (PTFE) was proposed as a means of simultaneously isolating the optical fiber from the corrosive environment and avoiding undesirable mechanical tensions on the FBGs. The presented sensor head is based on multiple FBGs inscribed in a lengthy single mode fiber. The sensor presents an average thermal sensitivity of 8.82 ± 0.09 pm/°C, resulting in a typical temperature resolution of ~0.1 °C and an average time constant value of 6.25 ± 0.08 s. Corrosion and degradation resistance were verified by infrared spectroscopy and scanning electron microscopy during 90 days exposure to high salinity crude oil samples. The developed sensor was tested in a field pilot test, mimicking the operation of an inland crude tank, demonstrating its abilities to dynamically monitor temperature profile.

Meeting optical QoS requirements with reduced complexity in dynamic wavelength assignment

Physical layer impairments may require highly complex implementations of the control plane in dynamic networks with quality of service (QoS) constraints. In order to alleviate this problem, we propose that dynamic wavelength assignment should be based on the elementary first-fit (FF) algorithm using off-line optimized priority lists. It is shown that this may significantly reduce the effects of physical impairments with no additional computational cost for the control of dynamic networks. A simple metric is proposed to evaluate optical QoS taking the performance of the network as a whole. Different heuristics are presented and evaluated. Results show that there are particular lists capable of reducing in four orders of magnitude the probability of FWM impacting network performance.

Multi-pump optimization for Raman+EDFA hybrid amplifiers under pump residual recycling

A new technique to optimize the gain profile of broadband hybrid amplifiers is here proposed. It is based on the equalization of Erbiums gain spectrum by adjusting the flexible gain profile of a multi-pump Raman amplification stage. However, for increased power pump efficiency, the residual Raman pump is recycled by the Erbium-doped section. This hybrid amplifier is assessed in terms of single-channel global gain, ripple, and noise figure. Ripple values below 1 dB have been found in a bandwidth of approximately 30 nm under two and three pump configurations.

Differentiated optical qos under a low complexity FWM-aware wavelength assignment algorithm

1 Abstract— Optical Quality of Service (QoS) for different classes of service in dynamic transparent optical networks is here addressed through Four Wave Mixing (FWM)-aware algorithms. The traffic classes differ in quality attributes, namely, transmission Bit Error Rate (BER) and blocking probability. The algorithms are tested on two mesh topologies against offered traffic, for different numbers of wavelengths per link. Results suggest that in order to simultaneously maintain Optical Service Level Agreement (OSLA) and efficient utilization of the resources, there must be some form of Connection Admission Control (CAC) mechanisms able to provide class isolation. This paper shows that this can be done either explicity or through a simple mechanism that is able to reduce lightpath non-linear interaction and favour the admission of high-priority traffic. A scalable and efficient mechanism based on shortest route, congestion avoidance and first fit wavelength assignment is suggested as a practical solution to QoS-Aware multiservice optical network.

Gain profile optimization for Raman+EDFA hybrid amplifiers with recycled pumps for WDM systems

An approximated technique to quickly optimize the gain profile of multi-pump broadband hybrid amplifiers under residual pump recycling in WDM systems is proposed. It is based on the amplifier spectral equalization by adjusting the flexible gain profile of a multi-pump Raman amplification stage to compensate the Erbium-doped stage gain variation within the C-band. In order to increase the pump power efficiency, the Erbium-doped fiber section recycles the residual Raman pump. Three different cases considering, respectively, one, two, and three pumps are analyzed to demonstrate the effectiveness of the proposed technique. The characterization of the Raman+EDFA hybrid amplifier has been performed in terms of global gain, ripple, and noise figure. The proposed optimization method for the Raman pump sets allows ripple figures to be reduced by half of their original values in single pumped configurations. Our results also suggest that two pump lasers is the best configuration when the maximum gain, minimum ripple, and pump power efficiency are simultaneously considered.

Simple ILP approaches to grooming, routing, and wavelength assignment in WDM mesh networks

This paper presents novel integer linear programming (ILP) formulations for integrated traffic grooming, routing (GR), and wavelength assignment (GRWA) designs. The objectives are i) to find out the minimal number of electro-optical transceivers needed for the whole network or ii) to design networks with the minimal number of transceivers at the node which requires the most, i.e., a min-max relaxation of the previous objective. Results are presented for mesh networks with 6, 10, and 16 nodes. The straightforward formulation proposed enables the minimal number of transceivers to be found below one minute running time for 6 nodes while relaxed approaches are able to solve a 16-node networks within one hour. However, it was found that min-max based designs are far from reaching the most economical solution.

FlexForward: Enabling an SDN manageable forwarding engine in Open vSwitch

This paper presents FlexForward, an approach for dynamically manage SDN data plane forwarding mechanisms. This feature leverages SDN flexibility in two aspects: (i) it provides a new management method, in which topological infrastructure requirements are satisfied by the most appropriate forwarding mechanism, residing in each managed network element; (ii) further decoupling between control and data planes in SDN by offering tableless forwarding support. As proof of concept, an implementation with different forwarding methods is carried out in Open vSwitch. Results show that FlexForward is able to achieve seamless switchover between forwarding methods. It also allows efficient tableless forwarding implementations, improving by up to 31% in latency, and 90% in throughput, when compared to regular OpenFlow.