Abstract The deployment of virtual network functions (VNFs) at edge servers potentially impairs the performance of latency-sensitive applications due to their computational cost. This work considers a new approach to addressing this problem that provides line rate acceleration of VNFs by employing field-programmable gate array (FPGA) equipped edge servers. This approach has been validated by practical use cases with both TCP and UDP as underlying protocol on a physical testbed environment. We examine the performance implications of executing a security VNF at an FPGA-equipped edge server. We experimentally demonstrate reduced VNF execution latency and energy consumption for a real-time video streaming application in comparison with a software-only baseline. In particular, the results show that the approach lowers VNF execution latency and power consumption at the edge by up to 44% and 76%, respectively, in our experiments while satisfying time constraints and maintaining confidentiality with high scalability. We also highlight the potential research challenges to make this approach viable in practice.

Energy efficient routers are important and promising devices in the roadmap toward green networking. In this paper, we explore this topic with a traffic aware design that can automatically adapt the router's energy consumption to the network traffic in real time with transitions at the scale from microseconds to nanoseconds. In this paper, we make three contributions. First, we propose an energy efficient router architecture using frequency scaling, which is extended from the contemporary line card architecture of the router. Second, we prototype our architecture on a gigabit NetFPGA platform, which extends a router design on that platform to dynamically work at six frequencies. Then, we discuss the implementation in detail and present a per-packet-based power consumption model from real-world measurements. Third, we further propose four algorithms, which efficiently and automatically adapt power grades under the methodology of periodical and threshold-based scaling. Our experiments are carried out on two platforms, a NetFPGA prototype for feasible scenarios and a simulator for other scenarios in excess of the NetFPGA port limit with parameters synthetically generated to the real-world measurement. Experimental results show that approximately 25% power can be saved with our architecture and algorithms. Conclusions are also given to summarize our findings.

Energy efficiency is now a significant concern for network infrastructure and next-generation network devices are expected to embed advanced power management capabilities. However, the effective exploitation of advanced power management capabilities in network devices which adaptively meet network load and operational constraints is still a considerable challenge due to the stochastic properties of the actual traffic load. In this paper, a statistical optimal local control policy for dynamic control of power state configurations according to the actual traffic load is proposed to minimize the power consumption while meeting the performance constraints. The packet-level statistical features of network traffic load is modeled as a first-order Markov chain and the dynamic power state selection problem is formulated as a Markov decision process, which can be solved using dynamic programming. In addition, we discuss the possibility of implementing the proposed scheme in real network devices, and design a case study in an NetFPGA frequency scaled router. Simulation results are presented to show the effectiveness of the proposed scheme.

Energy efficient routers are a promising step forward in the road-map towards energy efficient networking. In this paper, we use gigabit routers to measure and evaluate the impacts of green technologies in different scenarios. These programmable NetFPGA based routers come with the inbuilt feature of frequency scaling which can reduce the operational frequency of the board by half from 125MHz to 62.5MHz. Router power consumption is compared for standard and low frequency modes with different numbers of activated ports. Perpacket and per-byte energy costs in different cases with different packet sizes are measured and analyzed to evaluate the effects of frequency adjustment. Finally, a simple energy model for real traffic is proposed and evaluated. This paper reveals the internal power usage of a gigabit router at the granularity of packet, and header and payload byte level, and offers the reader a clear understanding of the potential power savings at different operational router frequencies, for different numbers of activated ports, various traffic loads and packet sizes, which would assist the green data center networks.

Rearrangeable switching networks have been a topic of interest for a long time among the research community. Routing algorithms for this class of networks have attracted lots of researchers along with other related areas, such as modification of the networks structure, crosspoints reduction etc. In this paper a new routing algorithm is presented for symmetric rearrangeable networks built with 2 × 2 switching element. A new matrix based abstraction model is derived to determine conflict free routing paths for each input-output request. Each stage of a network is mapped into a set of sub-matrices and number of matrices in each stage correspond to number of subnetworks in that stage. Once the input permutation is given, matrix cells are populated with binary values depending on the position of the switching elements in the actual hardware and their mapped matrix cells. These binary values control the routing decision in the underlying hardware. This new routing algorithm is capable of connection setup for partial permutation, m̅ = ρN, where N is the total input numbers and m̅ is the number of active inputs. Overall the serial time complexity of this method is O(NlogN)1and O(m̅logN) where all N inputs are active and with m̅2N). With m̅ active requests the time complexity goes down to O(logm̅logN), which is better than the O(log2m̅ + logN), reported in the 21/2[((log2N-4logN)1/2-logN)] ≤ ρ ≤ 1. Later half of this paper demonstrates how this routing algorithm is applicable for crosstalk free routing in optical domain.

The recent trend for Software as a Service and other types of cloud services is driving demand for data centers of ever increasing scale. This will require the scaling up and scaling out of existing data center architectures and will ultimately need new architectures featuring new topologies for the data center networks that interconnect its constituent nodes. The preferred tool for the detailed evaluation of such designs is simulation, as it gives finer detail than analysis at lower cost than test bed evaluation. Realistic simulation times require that the simulation itself be capable of being scaled out, i.e., it should be amenable to parallelization. We describe a simulation framework, together with a methodology for partitioning the relevant simulation models to allow their parallel implementation, and demonstrate its validity by applying it to the simulation of a hybrid optical/electrical network architecture using a cluster of high-end servers. We report on results capturing the performance of our simulator and discuss how these are associated with the underlying hardware hosting the simulation.

Despite much research effort key distribution in Wireless Sensor Networks (WSNs) still remains an open problem. In this paper we address this issue by proposing a simple identity- based key agreement scheme. Our protocol uses Identity-Based Cryptography (IBC) and secret key pre-distribution. We argue that IBC is in many ways a perfect solution for WSNs. It reduces the number of required keys, simplifying the key management in the network, and has a lower communications overhead than traditional public key protocols. We evaluate our proposal on a broad range of sensor platforms to show its efficiency on different CPU architectures. A complete key agreement procedure takes less than 3s on a resource- constrained Tmote Sky node without requiring any communication between two parties. To our knowledge this work is the first practical implementation of a complete IBC scheme on sensor devices. We identify a range of WSN applications which would benefit from the incorporation of a security architecture, and show that the scheme described here makes it feasible to deploy these applications in the real world.

Recent research results have shown that Elliptic Curve Cryptography (ECC) is feasible on resource constrained sensor nodes. In this work we demonstrate that the related but more complex primitives of Pairing Based Cryptography(PBC) are also well suited for sensor devices.We present the first in-depth study on the application and implementation of PBC to Wireless Sensor Networks (WSNs). Our implementations are all the fastest yet reported, and have been implemented across a range of WSN processors. On a system level we investigate the application of a simple non-interactive key exchange scheme that is particularly suitable for many WSN scenarios. We also present a novel variant of the key exchange protocol which can be useful in even more demanding applications, and which partially solves the problem of node compromise attacks.

The time synchronization problem needs to be considered in a distributed system. In Wireless Sensor Networks (WSNs) this issue must be solved with limited computational, communication and energy resources. Many synchronization protocols exist for WSNs. However, in most cases these protocols are independent entities with specific packets, communication scheme and network hierarchy. This solution is not energy efficient. Because it is very rare for synchronization not to be necessary in WSNs, we advocate integrating the synchronization service into the routing layer. We have implemented this approach in a new synchronization protocol called Routing Integrated Synchronization Service (RISS). Our tests show that RISS is very time and energy efficient and also is characterized by a small overhead. We have compared its performance experimentally to that of the FTSP synchronization protocol and it has proved to offer better time precision than the latter protocol.

High-end networking applications such as Internet TV and software distribution have generated a demand for multicast protocols as an integral part of the network. This will allow such applications to support data dissemination to large groups of users in a scalable and reliable manner. Existing IP multicast protocols lack these features and also require state storage in the core of the network which is costly to implement. In this paper, we present a new multicast protocol referred to as MENU. It realises a scalable and a reliable multicast protocol model by pushing the tree building complexity to the edges of the network, thereby eliminating processing and state storage in the core of the network. The MENU protocol builds multicast support in the network using mobile agent based active network services, Netlets, and unicast addresses. The multicast delivery tree in MENU is a two level hierarchical structure where users are partitioned into client communities based on geographical proximity. Each client community in the network is treated as a single virtual destination for traffic from the server. Netlet based services referred to as hot spot delegates (HSDs) are deployed by servers at "hot spots" close to each client community. They function as virtual traffic destinations for the traffic from the server and also act as virtual source nodes for all users in the community. The source node feeds data to these distributed HSDs which in turn forward data to all downstream users through a locally constructed traffic delivery tree. It is shown through simulations that the resulting system provides an efficient means to incrementally build a source customisable secured multicast protocol which is both scalable and reliable. Furthermore, results show that MENU employs minimal processing and reduced state information in networks when compared to existing IP multicast protocols.

It is said that a picture is worth a thousand words - this statement also applies to networking topics. Thus, to effectively monitor network performance we need tools which present the performance metrics in a graphical way which is also clear and informative. We propose a tool for this purpose which we call the traffic engineering eye diagram (TEED). Eye diagrams are used in digital communications to analyse the quality of a digital signal; the TEED can similarly he used in the traffic engineering field to analyse the load balancing ability of a TE algorithm. In this paper we describe how to create such TEEDs and how to use them to analyse and compare various traffic engineering approaches.

In a continually evolving Internet, tools such as quality of service routing must be used in order to accommodate user demands. QoS routing raises scalability issues within very large networks, which can he avoided by using hierarchical routing strategies. However, such strategies can lead to inaccurate path selection due to the aggregation process. To avoid such problems, we propose a hierarchical routing protocol, called macro-routing, which can distribute the route computation more efficiently throughout the network using mobile agents. It processes more detailed information than conventional hierarchical routing protocols, so is more likely to find the best path between source and destination. Also, by using mobile agents, more than one available path can be found. This provides a fast recovery mechanism, where no protocol restart is needed in a failure situation.

High-end networking applications such as e-commerce, multimedia, distributed data analysis and advanced collaborative environments feature demanding end-to-end quality of service (QoS) requirements. Due to the heterogeneity exhibited by the Internet, a route from source to destination for such a flow may not be available which is comprised exclusively of QoS supporting path segments. Hence the flow must traverse one or more non-QoS path segments referred to here as reservation gaps. In this paper we study the problem of reservation gaps and their impact on QoS and present a solution to address the deficiencies caused by such gaps, using an active network approach based on the mobile agent paradigm. Furthermore, to improve the reliability in path selection and to minimise the influence of reservation gaps along the path of a QoS flow, we propose two routing algorithms, the most reliable shortest path (MS-R) algorithm and the shortest - most reliable path (S-MR) algorithm, that select paths with the minimum number of reservation gaps. The active network based solution we propose works autonomously and scales to large networks such as the Internet. We demonstrate the advantages of such a solution using simulations which compares operational characteristics of QoS flows when traversing non-managed and actively managed reservation gaps. We also demonstrate the benefits of employing a routing algorithm such as MR-S or S-MR that accounts for reservation gaps in place of conventional shortest-path routing algorithms.

Multimedia traffic and real-time e-commerce applications can experience quality degradation in traditional networks such as the Internet. These difficulties can be overcome in networks which feature dynamically set up paths with bandwidth and delay guarantees. The problem of selecting such constrained paths is the task of quality of service (QoS) routing. This paper considers link-state routing, and the choice of cost metric used to implement QoS routing. There are two schools of thought regarding the choice of link cost. It is commonly assumed that QoS routing algorithms should limit hop count so as to conserve resources for future connections. Others advocate load balancing mechanisms so as to increase overall network utilisation. This paper investigates which of these approaches gives the better performance. We show that there is no one general answer to this question. We also point out the dangers of drawing general conclusions about routing algorithm performance based on the study of only a limited set of network topologies.

Multimedia traffic and real-time e-commerce applications can experience quality degradation in traditional networks such as the Internet. These difficulties can be overcome in networks which feature dynamically set up paths with bandwidth and delay guarantees. The problem of selecting such constrained paths is the task of quality of service (QoS) routing. Researchers have proposed several ways of implementing QoS routing, preferring either mechanisms which distribute network load or algorithms which conserve resources. Our previous studies have shown that network connectivity is an important factor when deciding which of these two approaches gives the best performance. In this paper we propose an algorithm, which features both load distribution and resource conservation. It takes a hybrid approach which balances between these two extreme approaches, according to the level of network connectivity. Our simulations indicate that this algorithm offers excellent performance over a than existing algorithms.

A method is described for performing routing in three-stage asynchronous transfer mode (ATM) switches which feature multiple channels between the switch modules in adjacent stages. The method is suited to hardware implementation using parallelism to achieve a very short execution time. This allows cell-level routing to be performed, whereby routes are updated in each time slot. The algorithm allows a contention-free routing to be performed, so that buffering is not required in the intermediate stage. An algorithm with this property, which preserves the cell sequence, is referred to as a path allocation algorithm. A detailed description of the necessary hardware is presented. This hardware uses a novel circuit to count the number of cells requesting each output module, it allocates a path through the intermediate stage of the switch to each cell, and it generates a routing tag for each cell, indicating the path assigned to it. The method of routing tag assignment described employs a nonblocking copy network. The use of highly parallel hardware reduces the clock rate required of the circuitry, for a given-switch size. The performance of ATM switches using this path allocation algorithm has been evaluated by simulation, and is described.

A method for path allocation for use with three-stage ATM switches that feature multiple channels between the switch modules in adjacent stages is described. The method is suited to hardware implementation using parallelism to achieve a very short execution time. This allows path allocation to be performed anew in each time slot. A detailed description of the necessary hardware is presented. This hardware counts the number of cells requesting each output module, allocates a path through the intermediate stage of the switch to each cell, and generates a routing tag for each cell, indicating the path assigned to it.>