Software switches are increasingly used in network function virtualization (NFV) to route traffic between virtualized network functions (VNFs) and physical network interface cards (NICs). Understanding of alternative switch designs remains deficient, however, in the absence of a comprehensive, comparative performance analysis. In this paper, we propose a methodology intended to be fair and use it to compare the performance of seven state-of-the-art software switches. We first explore their respective design spaces and then compare their performance under four representative test scenarios. Each scenario corresponds to a specific case of routing NFV traffic between NICs and/or VNFs. Our experimental results show that no single software switch prevails in all scenarios. It is therefore important to choose the one that is best adapted to a given use-case. The presented results and analysis bring a better understanding of design tradeoffs and identify potential bottlenecks that limit the performance of software switches.

Interest in software networking has grown significantly since the introduction of Network Function Virtualization (NFV). Software switches are used in NFV to steer traffic between different virtualized network functions and physical Network Interface Cards (NICs). It is becoming more and more important to objectively evaluate and compare the performance of the multiple alternative implementations that have recently been proposed. A comprehensive performance analysis is still missing for two main reasons: (i) the amount of time required to configure and compare all such tools is enormous; (ii) it is very difficult to define a proper methodology to compare different solutions in a fair manner. In this paper we propose a methodology based on four simple yet representative test scenarios used to evaluate the performance of software switches. We apply this methodology to measure throughput and latency metrics for 6 state-of-the-art software switches namely, OVS-DPDK, snabb, BESS, FastClick, VPP and netmap VALE. Our work constitutes a first step to building a better understanding of design tradeoffs and identifying performance bottlenecks.

Ever since Network Functions Virtualization has replaced dedicated appliances, ISPs have been able to add a degree of flexibility in their traffic engineering. However, it also has increased the complexity of the optimization problem, because it is now necessary to place virtual functions and route traffic jointly. Insofar, a logically centralized approach has been taken, where a so-called orchestrator, having full knowledge of the network, the virtual functions, and the traffic, run complex algorithms to find a suitable solution to the problem. The outcome of the algorithms are then translated to network configurations to be pushed to all of the appliances. We argue that there is no need to fully centralize every decision, rather we can leverage existing network intelligence to achieve the same goal. In particular we propose to augment the routing layer with the notion of services, so to rely on the robustness and scalability of Interior Gateway Protocols (IGP). Our solution leverages on existing distributed routing protocols where, in addition, autonomous nodes announce information about the virtual services they provide. Our design is modular and incrementally deployable and has been implemented in what we call a NFV Router. In our evaluation, we show that (i) NFV Routers distributed chaining decisions are close to optimal centrally-computed paths, (ii) on a large scale testbed deployment, NFV Routers efficiently steer traffic through chains and only add a small overhead to control traffic and (iii) our distributed system, because of its local control loop, has a faster reaction to network events than centralized solutions.

Ever since Network Functions Virtualization has replaced dedicated appliances, ISPs have been able to add a degree of flexibility in their traffic engineering. Now traffic is steered through virtual functions before reaching its destination, which, nonetheless, hardens network operation. Insofar, a logically centralized approach has been proposed to compute and install the correct forwarding behavior on every network devices. However, this flexibility comes at the cost of increased fragility compared to current Interior Gateway Protocols (IGP) and of a whole new infrastructure. Instead, we propose to augment the routing layer with the notion of services. Our solution leverages on existing distributed routing protocols where, in addition, autonomous nodes announce information about the virtual services they provide. Our design is modular, incrementally deployable and has been implemented in what we call an NFV Router. In our evaluation, we show that (i) NFV Routers distributed chaining decisions are close to optimal centrally-computed paths, (ii) on a large scale testbed deployment, NFV Routers efficiently steer traffic through chains and only add a small overhead to control traffic, and (iii) our distributed system, because of its local control loop, has a faster reaction to network events than centralized solutions.

Inter-domain Traffic Engineering for multi-homed networks faces a scalability challenge, as the size of BGP routing table continue to grow. In this context, the choice of the best path must be made potentially for each destination prefix, requiring all available paths to be characterized (e.g., through measurements) and compared with each other. Fortunately, it is well-known that a few number of prefixes could carry a dominant part of the traffic. As a natural consequence, to engineer a majority of traffic only a handful of prefixes need to be managed. Yet, traffic characteristics of a given prefix can vary greatly over time, which requires us to predict most significant prefixes in the near future. Moreover, little is known on the traffic dynamism at this aggregation level and sophisticated prediction methods won't scale in such context. In this paper, we study the relationship between prefix volume, stability, and predictability, based on recent traffic traces from nine different networks. Three simple and resource-efficient methods to select the prefixes associated with the most important foreseeable traffic volume are then proposed. Such proposed methods are capable of select sets of prefixes with both excellent representativeness (volume coverage) and stability in time, for which the best routes are identified. The analysis carried out confirms the potential benefits of a route decision engine.

The \emph{Locator/Identifier Separation Protocol} (LISP) introduces several benefits to the Internet architecture, yet, since it is just in the initial deployment stage, comprehensive understanding of its integration performance with legacy Internet becomes essential. We leverage on RIPE Atlas, the largest Internet measurement infrastructure, to conduct large scale measurements analysis to provide the feedback to improve LISP technology. The preliminary evaluations show that LISP generally has a reliable performance, compared with the existing Internet. From our vantage point, we observe that LISP introduces a non-negligible latency for the European and North American destinations, occasionally some extremely large delay, however, it shows a faster connection for the Asian intercontinental transmission.

In the SDN approach, the control and data planes are separated to enable programmability and interoperability, the control plane being logically centralized. Scalability and reliability are key properties for the SDN control plane. In WAN deployments, it is already envisioned as physically distributed to span across the networks and avoid single point of failure. In this paper, we tackle the problem of minimizing the overhead generated by distributed SDN controllers. We first model this problem as a Mixed Integer Program (MIP) and then propose a greedy algorithm. The results on the GEANT topology and on random graphs with up to 100 nodes show that our heuristic approaches to less than 3% the optimal in a manageable time.

The maturity of Software Defined Network (SDN) and network virtualization has inspired the development of software switch. This paper studies the behavior of soft switch in network virtualization environment. We investigate main performance factors and evaluate their impacts. Based on these measurements, we extend resource allocation among multiple soft switch instances with consideration of topology design. Simulation results show that proper coordination between resource allocation and topology greatly improves performance.

The future Internet has been a hot topic during the past decade and many approaches towards this future Internet, ranging from incremental evolution to complete clean slate ones, have been proposed. One of the proposition, LISP, advocates for the separation of the identifier and the locator roles of IP addresses to reduce BGP churn and BGP table size. Up to now, however, most studies concerning LISP have been theoretical and, in fact, little is known about the actual LISP deployment performance. In this paper, we fill this gap through measurement campaigns carried out on the LISP Beta Network. More precisely, we evaluate the performance of the two key components of the infrastructure: the control plane (i.e., the mapping system) and the interworking mechanism (i.e., communication between LISP and non-LISP sites). Our measurements highlight that performance offered by the LISP interworking infrastructure is strongly dependent on BGP routing policies. If we exclude misconfigured nodes, the mapping system typically provides reliable performance and relatively low median mapping resolution delays. Although the bias is not very important, control plane performance favors USA sites as a result of its larger LISP user base but also because European infrastructure appears to be less reliable.

In this paper, we propose a two-layered parking lot management system for charging scheduling of electric vehicles (EVs) considering a realistic vehicular mobility pattern. EVs are categorized in two groups based on their mobility patterns: Regular EVs and Irregular EVs. We use the data from an vehicular mobility trace collected from the Canton of Zurich for the regular EVs and a probabilistic pattern built on top of this Zurich trace aiming at modeling the behavior of irregular EVs. To the extend of our knowledge, this is the first EV charging scheduling study in the literature that utilizes a big-scale realistic vehicular mobility trace. The performance of the proposed system is compared with well-known routine unaware scheduling mechanisms (e.g., First Come First Serve) with regard to maximizing the parking lot revenue. Our results show that, our proposed system outperforms well-known routine unaware scheduling mechanisms and it is evident that real environments would benefit from using such parking lot management systems in Smart Cities.

Ontologies are being developed for the life sciences to standardise the way we describe and interpret the wealth of data currently being generated. As more ontology based applications begin to emerge, tools are required that enable domain experts to contribute their knowledge to the growing pool of ontologies. There are many barriers that prevent domain experts engaging in the ontology development process and novel tools are needed to break down these barriers to engage a wider community of scientists. We present Populous, a tool for gathering content with which to construct an ontology. Domain experts need to add content, that is often repetitive in its form, but without having to tackle the underlying ontological representation. Populous presents users with a table based form in which columns are constrained to take values from particular ontologies. Populated tables are mapped to patterns that can then be used to automatically generate the ontology's content. These forms can be exported as spreadsheets, providing an interface that is much more familiar to many biologists. Populous's contribution is in the knowledge gathering stage of ontology development; it separates knowledge gathering from the conceptualisation and axiomatisation, as well as separating the user from the standard ontology authoring environments. Populous is by no means a replacement for standard ontology editing tools, but instead provides a useful platform for engaging a wider community of scientists in the mass production of ontology content.

SNOMED-CT is a large medical terminology based on description logic and mandated for use in the US, UK and several other countries. The hierarchies are known to contain many errors, but have so far proved difficult to analyse or quality assure. We present a series of methods and lessons learnt from experience in quality assuring a "module" of SNOMED for specific applications that we expect to generalize both to SNOMED as a whole and to other large ontologies. They feature a) dependence on domain exper-tise b) starting from classes selected for relevance to specific applications, c) tracing all errors to their root and verifying repairs by reclassification d) extraction of manageable-sized "modules"; e) mixed semantic and lexical criteria, and f) extensive use of scripting. They aim to reduce the cognitive load on experts by a) looking initially up-wards rather than downwards in the hierarchies, b) breaking up long lists of direct subclasses by introducing definitions for meaningful subcategories. Errors found range from simple mistakes to systematic errors in schemas. © 2011 ACM.

Recent activities in the IRTF (Internet Research Task Force), and in particular in the Routing Research Group (RRG), focus on defining a new Internet architecture, in order to solve scalability issues related to interdomain routing. The research community has agreed that the separation of the end-systems' addressing space (the identifiers) and the routing locators' space will alleviate the routing burden of the Default Free Zone. Nevertheless, such approach, adding a new level of indirection, implies the need of storing and distributing mappings between identifiers and routing locators. In this paper we present LISP-DHT, a mapping distribution system based on Distributed Hash Tables (DHTs). LISP-DHT is designed to take full advantage of the DHT architecture in order to build an efficient and secured mapping lookup system while preserving the locality of the mapping. The paper describes the overall architecture of LISP-DHT, explaining its main points and how it works.

Keyword-based search engines, though hugely popular, show limitations when trying to answer very specific queries. The processing of search results is performed by users, rather than by software. Ontologies provide a means to create formal, machineprocessable descriptions of the knowledge in a certain domain [14], and, by using elements of these descriptions to annotate suitable information sources, they can be analysed and manipulated in an intelligent manner. The QuestSemantics platform provides automated ontology-based metadata creation and resource annotation, with subsequent ontology-based querying of the annotated resources. The platform has been deployed in two commercial scenarios, providing useful feedback on both the feasibility and effectiveness of applying Semantic Web technologies to specific business problems.

In this paper we therefore propose a reformulation of the entropy metric to evaluate the amount of information carried by both the ontology structure, and also by the language elements (i.e. the semantics associated with the edges in the ontological graph). To evaluate this approach, the reformulated metric is empirically compared to Calemt & Daemi's original entropy metric, for a variety of different sized modules. The results suggest that not only can entropy differentiate between structurally different modules of the same size, but that our improved entropy metric provides a finer grain differentiation than the original entropy metric.

Changes in an ontology may have a disruptive impact on any system using it. This impact may depend on structural changes such as introduction or removal of concept definitions, or it may be related to a change in the expected performance of the reasoning tasks. As the number of systems using ontologies is expected to increase, and given the open nature of the semantic Web, introduction of new ontologies and modifications to existing ones are to be expected. Dynamically handling such changes, without requiring human intervention, becomes crucial. This paper presents a framework that isolates groups of related axioms in an OWL ontology, so that a change in one or more axioms can be automatically localised to a part of the ontology.

Problems resulting from the management of shared, distributed knowledge has led to ontologies being employed as a solution, in order to effectively integrate information across applications. This is dependent on having ways to share and reuse existing ontologies; with the increased availability of ontologies on the web, some of which include thousands of concepts, novel and more efficient methods for reuse are being devised. One possible way to achieve efficient ontology reuse is through the process of ontology module extraction. A novel approach to ontology module extraction is presented that aims to achieve more efficient reuse of very large ontologies; the motivation is drawn from an Ontology Engineering perspective. This paper provides a definition of ontology modules from the reuse perspective and an approach to module extraction based on such a definition. An abstract graph model for module extraction has been defined, along with a module extraction algorithm. The novel contribution of this paper is a module extraction algorithm that is independent of the language in which the ontology is expressed. This has been implemented in ModTool; a tool that produces ontology modules via extraction. Experiments were conducted to compare ModTool to other modularisation methods.

Very recent activities in the IETF and in the Routing Research Group (RRG) of the IRTG focus on defining a new Internet architecture, in order to solve scalability issues related to interdo-main routing. The approach that is being explored is based on the separation of the end-systems' addressing space (the identifiers) and the routing locators' space. This separation is meant to alleviate the routing burden of the Default Free Zone, but it implies the need of distributing and storing mappings between identifiers and locators on caches placed on routers. In this paper we evaluate the cost of maintaining these caches when the distribution mechanism is based on a pull model. Taking as a reference the LISP protocol, we base our evaluation on real Netflow traces collected on the border router of our campus network. We thoroughly analyze the impact of the locator/ID separation, and related cost, showing that there is a trade-off between the dynamism of the mapping distribution protocol, the demand in terms of bandwidth, and the size of the caches.

The success of the IEEE 802.11 standard has pushed the research community to put a not negligible effort in the design, analysis, and development of new wireless solutions during the last decade. More recently, the apparition of the wireless mesh technology incited a number of wireless testbeds to be deployed in several countries, in both academic and industrial environments. The LIP6 laboratory of the Université Pierre et Marie Curie is a main actor in the context of wireless mesh networking and has deployed its own testbed. Custom wireless routers built up from off-the-shelf hardware and open source software form this testbed, which is called MeshDVNet. On this platform the LIP6 has designed and implemented the MeshDV demon, an integrated approach for routing and mobility management, which hides to end-users all the complexity of wireless mesh networks. MeshDV allows end-users to easily access the network without the constrain of embedding any kind of additional software (e.g. routing demon, protocol stack extensions).

In the widely used 802.11 standard, the so-called performance anomaly is a well-known issue. Several works have tried to solve this problem by introducing mechanisms such as packet fragmentation, backoff adaptation, or packet aggregation during a fixed time interval. In this paper, we propose a novel approach solving the performance anomaly problem by packet aggregation using a dynamic time interval, which depends on the busy time of the wireless medium. Our solution differs from other proposition in the literature because of this dynamic time interval, which allows increasing fairness, reactivity, and in some cases efficiency. In this article, we emphasize the performance evaluation of our proposal.

Wireless Mesh Networks (WMNs) are an emerging architecture based on multi-hop transmission. ISPs considers WMNs as a potential future technology to offer broadband Internet access. Therefore, in WMNs, the throughput capacity of the wireless backbone becomes a key factor, limiting the scalability in terms of users able to effectively take advantage of the network. Increasing the effective throughput capacity will support WMNs become a very cost-effective solution for wireless ISPs.Based on the analysis of the theoretical capacity bound derived by Gupta et al., we propose MRS (Mesh Routing Strategy), a novel cross layer routing solution. MRS is specifically designed for Wireless Mesh Networks, where routing is challenging due to the unreliable wireless medium. Traditional routing paradigms are not able to overcome this issue, resulting in an average throughput experienced by the network far lower than the theoretical throughput capacity bound. This paper shows how a cross-layer approach will consistently improve the throughput experienced by large-scale multi-hop networks. We show through simulation that our proposal increases throughput while reducing the average transmitted power. Furthermore, MRS limits interference and gets closer to the theoretical throughput capacity bound.

We focus on the problem of specialization in a description logics (DL) representation, specifically the ALN language. Standard approaches to learning in these representations are based on bottom-up algorithms that employ the lcs operator, which, in turn, produces overly specific (overfitting,) and still redundant concept definitions. In the dual (top-down) perspective, this issue can be tackled by means of an ILP downward operator. Indeed, using a mapping from DL descriptions onto a clausal representation, we define a specialization operator computing maximal specializations of a concept description on the grounds of the available positive and negative examples.