The growth of data traffic on the web, the virtualization of services, and the changes in the pattern of traffic between users and data centers have led to a reassessment of the current methods of doing network administration. Software Defined Networks (SDNs) propose a paradigm that delegate the control of packets and flows to applications, developed according to specific requirements, where the OpenFlow protocol can be used for communications. The development of this type of applications, as in any other development area, requires tests and measurement tools to facilitate a performance evaluation. However, the current open-source performance measurement applications for SDN networks cover only very basic characteristics, while there is a wide range of SDN controllers with support to many versions of OpenFlow, making the selection of the controller a difficult point to address. In this paper, we propose a distributed performance evaluation tool for SDN controllers, that can assess the throughput, latency, percentage of memory consumption, percentage of CPU utilization, and consumption in kB for input/output interfaces, using OpenFlow version 1.3. Our tool is based on Cbench, and adds new functionalities such as the graphical representation of results to analyze the outcomes. To validate our tool, we make a performance evaluation of well-known SDN controllers such as Ryu, OpenDaylight, OpenMUL, and Floodlight, in environments under great stress of requests.

The current communication requirements are moving quickly to mobile ad-hoc networks, mainly because of the decreasing cost of wireless networking technology and the increasing capacity to integrate into embedded devices technologies such as WiFi, ZigBee, Bluetooth, and WAVE. It is important to understand the limitation or the strength of each protocol and standard in order to make a better selection decision at the moment of implementing a specific scenario. This paper evaluates how routing protocols perform in mobile scenarios when the speed of nodes is relatively high. Our research shows that reactive routing protocols have a better packet delivery ratio (PDR) than proactive routing protocols. However, proactive routing protocols have a smaller end-to-end delay and hops count.

InfiniBand is widely accepted as a high performance networking technology for datacenters and HPC clusters. It uses the Remote Direct Memory Access (RDMA) where communication tasks that are typically assigned to CPUs are offloaded to the Channel Adapters (CAs), resulting in a significant increase of the throughput and reduction of the latency and CPU load. In this paper, we make an introduction to InfiniBand and IP over InfiniBand (IPoIB), where the latter is a protocol proposed by the IETF to run traditional socket-oriented applications on top of InfiniBand networks. We also evaluate the performance of InfiniBand using different transport protocols with several benchmarking tools in a testbed. For RDMA communications, we consider three transport services: (1) Reliable Connection, (2) Unreliable Connection, and (3) Unreliable Datagram. For UDP and TCP, we use IPoIB. Our results show significant differences between RDMA and IPoIB communications, encouraging the coding of new applications with InfiniBand verbs. Also, it is noticeable that IPoIB in datagram mode and in connected mode have similar performance for small UDP and TCP payload. However, the differences get important as the payload size increases.

The exhaustion of IPv4 addresses has forced the deployment of the new version of the Internet Protocol (IPv6). However, the migration to the new protocol is done gradually and with the due care for many reasons that include: cost, inclusion of support for IPv6 in existing applications, training of technical staff, lack of web content available over IPv6 from important providers, and obsolete devices not anymore supported by manufacturers. For those reasons, many transition mechanisms have been proposed, each one to fill distinct requirements, with different operational theory and availability according to the network environment. A performance evaluation of these mechanisms can help network administrators and researchers in their selection of the best transition technology for their environment. In this paper, we present a performance comparison of some transition mechanisms such as ISATAP, 6to4, and NAT64 in real testbeds with Debian, Windows 7, Windows 8, and Windows 10. For NAT64, two different tools were tested: TAYGA and Jool. We measure the OWD and the throughput for UDP and TCP for every mechanism, for both Ethernet and Fast Ethernet technologies. From this research, we can conclude that all the modern operating systems for PCs already have good support for IPv6, and a very similar network performance. Also, we can infer from our work that in controlled environments, native IPv4 has the best performance, closely followed by native IPv6. The difference is essentially due to the length of the IP header (20 bytes in IPv4 and 40 bytes in IPv6). The tunneling solutions chosen for this research (ISATAP and 6to4) have a similar performance, which is the lowest of the studied technologies, because of the additional IPv4 header in the tunnel.