The Transport layer of OSI and TCP/IP models provides all necessary services for end-to-end commu... more The Transport layer of OSI and TCP/IP models provides all necessary services for end-to-end communication between application processes. There are a huge amount of works and propositions of Transport level protocols and services to satisfy applications requirements. Unfortunately, the vast majority of applications refer only to TCP for reliable and ordered services or to UDP for unreliable and low latency services. This is due to the fact that the deployment of all new Transport protocol proposal is mainly hampered by (1) the poor socket API exposed by the Transport layer to applications, (2) the introduction of middleboxes within the Internet and (3) the tedious work required to modified Operating System kernel. At the same time, the development of network functions virtualization opportunities is growing in the world of carrier networks and more generally. In this paper, after a survey on Transport protocols deployment issues, we present a novel approach to realize the effective deployment of Transport protocols and services by leveraging virtualization principles. At last, we present a new way to efficiently manage Transport functions and to dynamically build Transport services through a graph-based protocol model.
HAL (Le Centre pour la Communication Scientifique Directe), Dec 1, 2015
OneM2M standard is a global initiative led jointly by major standardization organizations around ... more OneM2M standard is a global initiative led jointly by major standardization organizations around the world in order to come up with a unique standard for M2M communications. Prior standards, but also oneM2M, while focusing on achieving interoperability at the communication level, fail or lack to achieve full interoperability at the semantic data level. An expressive ontology for IoT called IoT-O has been defined making best use of already defined ontologies in specific domains such as sensor, observation, service, quantity kind, units, or time. IoT-O defines also some missing concepts relevant for IoT such as thing, actuator, actuation, or manager. The extension of the oneM2M standard to support semantic data interoperability based on IoT-O is discussed. Finally, through comprehensive real use cases, benefits of the augmented standard are demonstrated ranging from heterogeneous devices interoperability to autonomic behavior achieved by automated reasoning.
HAL (Le Centre pour la Communication Scientifique Directe), Nov 13, 2017
The Internet of objects (IoT) will have to meet the non-functional needs (QoS, security, etc.) of... more The Internet of objects (IoT) will have to meet the non-functional needs (QoS, security, etc.) of new business applications supported by the cloud. To do this, the interactions between the underlying application software and the communicating objects will rely on networks and communication middleware with configurable, programmable and dynamically deployable capabilities. These capabilities will be available both on pre-existing entities but also on virtual entities, i.e. that will be dynamically created in the Cloud according to the need. In this new ecosystem, meeting the end-to-end QoS needs of these future applications is a major challenge. This challenge has particularly to be tackled both at the level of the Middleware intermediary entities and at the level of the networks interconnecting these entities. In this context, this paper presents our approach for a self-adaptive QoS management at the middleware level for IoT applications. This approach is aimed at: 1) taking advantage of the technological opportunities offered by the Cloud, the dynamic deployment of processing functions and the autonomic computing paradigm, 2) taking into account the heterogeneity of the solutions that will coexist in this landscape, and 3) ensuring the consistency of the (re) configuration choices thanks to appropriate theoretical tools.
The Internet of objects (IoT) will have to meet the non-functional needs (QoS, security, etc.) of... more The Internet of objects (IoT) will have to meet the non-functional needs (QoS, security, etc.) of new business applications supported by the cloud. To do this, the interactions between the underlying application software and the communicating objects will rely on networks and communication middleware with configurable, programmable and dynamically deployable capabilities. These capabilities will be available both on pre-existing entities but also on virtual entities, i.e. that will be dynamically created in the Cloud according to the need. In this new ecosystem, meeting the end-to-end QoS needs of these future applications is a major challenge. This challenge has particularly to be tackled both at the level of the Middleware intermediary entities and at the level of the networks interconnecting these entities. In this context, this paper presents our approach for a self-adaptive QoS management at the middleware level for IoT applications. This approach is aimed at: 1) taking advantage of the technological opportunities offered by the Cloud, the dynamic deployment of processing functions and the autonomic computing paradigm, 2) taking into account the heterogeneity of the solutions that will coexist in this landscape, and 3) ensuring the consistency of the (re) configuration choices thanks to appropriate theoretical tools.
The Internet of things (IoT) has evolved exceptionally in recent years; enabling a large number o... more The Internet of things (IoT) has evolved exceptionally in recent years; enabling a large number of heterogeneous devices to be interconnected to users via the Internet. This new concept promises in a few years to interconnect billions of devices, which will generate many challenges on the infrastructure supporting these communications. One of these challenges is the satisfaction of the different QoS requirements of the applications. To address this challenge, we identified two bottlenecks with respect to the QoS, which are the networks and the intermediate entities (i.e. middleware) allowing the applications to interact with the devices. In this paper, we propose a modular framework to ensure the QoS of applications at the middleware-level through QoS-oriented mechanisms deployed dynamically and autonomously on the middleware entities. The benefits of this framework are presented through test scenarios in the vehicular transportation domain.
IEEE Transactions on Network and Service Management, Dec 1, 2022
In this paper, we consider the problem of identifying multiple bottlenecks (a.k.a bottleneck anal... more In this paper, we consider the problem of identifying multiple bottlenecks (a.k.a bottleneck analysis) in IoT Service Platforms. For QoS-constrained applications, IoT Platforms have grown in complexity with non-stationary workloads and intertask dependencies created by data flows crossing the platform's nodes. These factors create multiple simultaneous "bottlenecks" (a bottleneck expresses overload in terms of request processing time on a given node, and contributes to QoS degradation). Multi-bottlenecks are non-trivial to analyze since they may escape typical assumptions made in classic performance analysis, such as analysis based on queuing theory models. Solving this analysis problem requires real-time collection and analysis of data that can be massive, and as a result, induce negative impacts on the performance of the NFV-based IoT Platform (NIP) (e.g., use of bandwidth, computing resource, and storage resource). Therefore, it needs to be adapted to the strict minimum allowing effective analysis. We build an adaptive performance analysis method that optimizes bottlenecks' identification for a monitoring overhead budget associated with the different available metrics. Instead of systematically collecting all the NIP metrics, the proposed process determines the best subset of metrics to consider for the efficiency of the performance analysis. The conducted experiments on a practical use case show that the proposed method exhibited high performances of the bottleneck analysis process, in the presence of different bottleneck types and durations, with very few false positives and false negatives.
Service-oriented computing and applications, Jun 26, 2019
The IoT platforms must allow the communication between the Applications and Devices according to ... more The IoT platforms must allow the communication between the Applications and Devices according to their non-functional requirements. One of the main non-functional requirements is the Quality of Service (QoS). In a previous work has been defined an Autonomic Internet of Things (IoT) platform for the QoS Management, based on the concept of autonomic cycle of data analysis tasks. In this platform have been defined two autonomic cycles, one based on a classification task that determines the current operational state to define the set of tasks to execute in the communication system to guarantee a given QoS. The other one is based on a clustering task that discovers the current operational state, and based on it, determines the set of tasks to be executed in the communication system. This paper analyzes the diagnostic capabilities of the system based on both approaches, using different metrics. For that, a real scenario has been considered, with simulations that have generated data to test both tasks. Each technique has different aspects to be considered for a correct QoS management in the context of IoT platforms. The classification technique can determine very well the learned operational states, but the clustering approach can carry out a more detailed description of the operational states. Additionally, due to the classification and clustering technique used, called LAMDA (Learning Algorithm for Multivariate Data Analysis), the paper analyzes the operational state profile determined by them, which is very useful in a diagnostic process.
HAL (Le Centre pour la Communication Scientifique Directe), Jul 19, 2021
Internet and its evolutions are fundamentally based on the unique TCP/IP model, whose primary pro... more Internet and its evolutions are fundamentally based on the unique TCP/IP model, whose primary protocol of the Transport layer (L4) is the TCP (and somewhat UDP). Despite its well-known limitations, TCP is still widely deployed and used on the Internet. Nearly all the literature's propositions to overcome TCP's limitations are not deployed in the mainstream operating systems (OS) and/or face limited use by the Internet's applications. This situation leads to the ossification or sclerosis of the Transport layer that is a significant barrier to the introduction of innovations in this layer of the Internet's TCP/IP architecture. Thus, this paper proposes to address this issue and focuses on two main contributions. First, we design and implement a Virtual Transport Layer (VTL) system able to dynamically deploy Transport protocols within the end-systems' OS. Second, to facilitate the use and stimulate the adoption of the proposed architecture (and the L4 protocols it deploys), we provide the approach and mechanisms necessary to allow TCPbased applications to use transparently any Transport protocol other than TCP. Experiments show the correctness of VTL and its ability to quickly deploy protocols. Further, we show that the deployed protocols achieve significant performances under VTL.
HAL (Le Centre pour la Communication Scientifique Directe), Feb 27, 2020
Internet and its evolutions are fundamentally based on the unique TCP/IP model, whose primary pro... more Internet and its evolutions are fundamentally based on the unique TCP/IP model, whose primary protocol of the Transport layer (L4) is the TCP (and somewhat UDP). Despite its well-known limitations, TCP is still widely deployed and used on the Internet. Nearly all the literature's propositions to overcome TCP's limitations are not deployed in the mainstream operating systems (OS) and/or face limited use by the Internet's applications. This situation leads to the ossification or sclerosis of the Transport layer that is a significant barrier to the introduction of innovations in this layer of the Internet's TCP/IP architecture. Thus, this paper proposes to address this issue and focuses on two main contributions. First, we design and implement a Virtual Transport Layer (VTL) system able to dynamically deploy Transport protocols within the end-systems' OS. Second, to facilitate the use and stimulate the adoption of the proposed architecture (and the L4 protocols it deploys), we provide the approach and mechanisms necessary to allow TCPbased applications to use transparently any Transport protocol other than TCP. Experiments show the correctness of VTL and its ability to quickly deploy protocols. Further, we show that the deployed protocols achieve significant performances under VTL.
HAL (Le Centre pour la Communication Scientifique Directe), Oct 1, 2014
Ce document a pour objectif de dresser un état de l'art des paradigmes réseaux émergents que nous... more Ce document a pour objectif de dresser un état de l'art des paradigmes réseaux émergents que nous envisageons de considérer dans le contexte du projet ADN, à savoir la virtualisation réseau, les réseaux programmables ou « Software-Defined Networks » (SDN), la virtualisation des fonctions réseau ou « Network Function Virtualization » (NFV) ainsi que le réseau dans le nuage ou « Cloud Networking ». Dans chacune des parties sont abordés : les principes généraux, un état des lieux des principales technologies ainsi que des cas d'utilisation représentatifs. Ce document introduit également le standard DDS (Data Distribution Service) de l'OMG (Object Management Group), middleware adopté dans le cadre de ce projet.
HAL (Le Centre pour la Communication Scientifique Directe), Nov 4, 2015
Le contenu de ce document est la propriété d'Airbus Defence and Space SAS. Il ne doit pas être di... more Le contenu de ce document est la propriété d'Airbus Defence and Space SAS. Il ne doit pas être diffusé à d'autres personnes que les destinataires. Sa reproduction non autorisée est interdite. Tous droits réservés.
Maturity of virtualization techniques and the success of rich cloud services have pushed the hori... more Maturity of virtualization techniques and the success of rich cloud services have pushed the horizon of a new sophisticated and connected IoT systems over the Internet. However, provisioning, including managing, such systems is still static, and consequently, costly and time consuming. It is poorly suited to the dynamic criterion of such systems. In this paper, self-adaptive management procedures are proposed towards QoS maintaining at the middleware level of IoT systems. The design is based on concepts such as Network Function Virtualization (NFV) and its generalization, Software-defined Networks (SDN) and edge computing. A realistic use case is implemented for illustration and validation purposes. Prototype execution shows the self-adaptive capabilities of the approach to maintain operative QoS while facing unpredicted events such as fire detection.
Resource monitoring is a key task in network management. The concept of Software Defined Networki... more Resource monitoring is a key task in network management. The concept of Software Defined Networking (SDN) allows taking benefits of the advantages of both active and passive monitoring techniques. However, this monitoring has a cost, hence the importance of the selection of the "key" switches to be interrogated and their polling frequency in order to reduce monitoring cost. This cost is expressed here in term of computing time. Monitoring of links can be used to determine the available bandwidth on each link, with the aim to meet the applicative QoS requirements based on appropriate routing. In this context, this paper first provides a formulation of the problem of choosing key switches as a vertex cover problem and proposes a heuristic method to solve the formulated problem. It then provides an implementation and a performance evaluation of the proposed algorithm within the Floodlight SDN controller. These performances are compared to those of the currently existing Floodlight monitoring module. Finally, we present one application of our proposed monitoring.
Fog Computing infrastructures are deployed in the immediate vicinity of users in order to meet th... more Fog Computing infrastructures are deployed in the immediate vicinity of users in order to meet the stringent delay requirements of some emerging IoT applications, which cannot be achieved with traditional Cloud Computing infrastructures. The latency gains of Fog Computing come however at the cost of a potentially larger total capacity. The duplication of ressources in many micro data centres may also lead to an explosion of energy and operations costs. In this paper, we consider the optimal capacity planning of Fog Computing infrastructures under probabilistic delay guarantees. Despite the non-linearity of the delay constraints, we show that the problem can be formulated as a Mixed Integer Linear Programming (MILP) problem. We first present a MILP formulation of the problem assuming that the infrastructure cost depends linearly on the capacities. To account for economies of scale in favour of large data centres, we then extend this MILP formulation to arbitrary concave objective functions. Empirical results show that the optimal capacity-planning solution can be determined efficiently even for large-size problem instances, and that it can results in significant gains with respect to the solution in which user requests are always processed in the nearest data centre.
In the recent years, telecom and computer networks have witnessed new concepts and technologies t... more In the recent years, telecom and computer networks have witnessed new concepts and technologies through Network Function Virtualization (NFV) and Software-Defined Networking (SDN). SDN, which allows applications to have a control over the network, and NFV, which allows deploying network functions in virtualized environments, are two paradigms that are increasingly used for the Internet of Things (IoT). This Internet (IoT) brings the promise to interconnect billions of devices in the next few years rises several scientific challenges in particular those of the satisfaction of the quality of service (QoS) required by the IoT applications. In order to address this problem, we have identified two bottlenecks with respect to the QoS: the traversed networks and the intermediate entities that allows the application to interact with the IoT devices. In this paper, we first present an innovative vision of a "network function" with respect to their deployment and runtime environment. Then, we describe our general approach of a solution that consists in the dynamic, autonomous, and seamless deployment of QoS management mechanisms. We also describe the requirements for the implementation of such approach. Finally, we present a redirection mechanism, implemented as a network function, allowing the seamless control of the data path of a given middleware traffic. This mechanism is assessed through a use case related to vehicular transportation.
IEEE Journal on Selected Areas in Communications, Apr 1, 2014
This paper addresses the problem of quality bottleneck in adaptive SVC streaming. Quality bottlen... more This paper addresses the problem of quality bottleneck in adaptive SVC streaming. Quality bottleneck occurs in adaptive streaming systems when the desired video quality cannot be obtained even if the network capabilities are sufficient. In the context of SVC layered video streaming, we have observed that enhancement layers remain around the video source and fail to reach all the participating peers. To overcome this problem, we propose an adaptive SVC streaming solution that cooperatively integrates strategies of overlay formation, data scheduling and content adaptation. Performance evaluation using simulations shows that the proposed streaming solution reduces the quality bottleneck, increases churn-tolerance and optimizes bandwidth utilization.
Sustainable Computing: Informatics and Systems, Dec 1, 2014
Best known Cloud providers propose services under constraints of Service Level Agreement (SLA) de... more Best known Cloud providers propose services under constraints of Service Level Agreement (SLA) definitions. The SLAs are composed of different Quality of Service (QoS) rules promised by the provider. Thus, the QoS in Clouds becomes more and more important. Precise definitions and metrics have to be explained. This article proposes an overview of Cloud QoS parameters as well as their classification, but also it defines usable metrics to evaluate QoS parameters. Moreover, the defined QoS metrics are measurable and reusable in any scheduling approach for Clouds. Energy consumption is an inherent objective in Cloud Computing, thus, it is also considered. For evaluation purposes, two uncommon QoS parameters: Dynamism and Robustness are taken into account in different Cloud virtual machines scheduling approaches. Validation is done through comparison of common scheduling algorithms, including a genetic algorithm (GA), in terms of QoS parameters evolution in time. Simulation results have shown that including various QoS parameters allows a deeper scheduling algorithms analysis.
Uploads
Papers by Samir Medjiah