DRAGON: a framework for service provisioning in heterogeneous grid networks

Dynamic resource allocation in GMPLS optical networks (DRAGON) defines a research and experimental framework for high-performance networks required by grid computing and e-science applications. The DRAGON project is developing technology and deploying network infrastructure which allows dynamic provisioning of network resources in order to establish deterministic paths in direct response to end-user requests. This includes multidomain provisioning of traffic-engineering paths using a distributed control plane across heterogeneous network technologies while including mechanisms for authentication, authorization, accounting (AAA), and scheduling. A reference implementation of this framework has been instantiated in the Washington, DC area and is being utilized to conduct research and development into the deployment of optical networks technologies toward the satisfaction of very-high-performance science application requirements.     

The OptIPuter: high-performance, QoS-guaranteed network service for emerging E-science applications

Emerging large-scale scientific applications have a critical need for high bandwidth and predictable-performance network service. The OptlPuter project is pioneering a radical new type of distributed application paradigm that exploits dedicated optical circuits to tightly couple geographically dispersed resources. These private optical paths are set up on demand and combined with end resources to form a distributed virtual computer (DVC). The DVC provides high-quality dedicated network service to applications. In this article we compare the OptIPuter's approach (DVC), which exploits network resources to deliver higher-quality network services, to several alternative service models (intelligent network and asynchronous file transfer). Our simulations show that there are significant differences among the models in their utilization of resources and delivered application services. Key takeaways include that the OptlPuter approach provides applications with superior network service (as needed by emerging e-science applications and performance-critical distributed applications), at an expense in network resource consumption. The other approaches use fewer network resources, but provide lower-quality application service.     

Utility-Aware Cognitive Network Selections in Wireless Infrastructures

Operators of wireless infrastructures should maintain their users “always-best-connected”. This concept means that applications should be offered to users at the best possible Quality of Service (QoS) level, taking into account profile, context and policy information. The profiles provide the user requirements and preferences, the terminal capabilities, and the application requirements. The policies provide the objectives, constraints imposed by various stakeholders, for instance the network operator (NO). The context of operation designates relevant applications, available networks and their QoS capabilities. The “always-best-connectivity” concept can be achieved by directing user terminals to the most appropriate networks of the heterogeneous infrastructure of the NO. In this respect, advanced terminal management functionality is required. This paper presents management mechanisms for utility-based cognitive network selections. The utility is used for expressing the user desire for a QoS level. Cognition mechanisms are applied for learning the QoS capabilities of candidate networks, and therefore increasing the reliability and seamlessness of the network selections. Extensive results are provided, which show the behaviour of the scheme in terms of network selections made, and computational effort required for the acquisition of the knowledge.     

Future Optical Networks

This paper presents views on the future of optical networking. A historical look at the emergence of optical networking is first taken, followed by a discussion on the drivers pushing for a new and pervasive network, which is based on photonics and can satisfy the needs of a broadening base of residential, business, and scientific users. Regional plans and targets for optical networking are reviewed to understand which current approaches are judged important. Today, two thrusts are driving separate optical network infrastructure models, namely 1) the need by nations to provide a ubiquitous network infrastructure to support all the future services and telecommunication needs of residential and business users and 2) increasing demands by the scientific community for networks to support their requirements with respect to large-scale data transport and processing. This paper discusses these network models together with the key enabling technologies currently being considered for future implementation, including optical circuit, burst and packet switching, and optical code-division multiplexing. Critical subsystem functionalities are also reviewed. The discussion considers how these separate models might eventually merge to form a global optical network infrastructure     

Optical dynamic intelligent network services (ODIN): an experimental control-plane architecture for high-performance distributed environments based on dynamic lightpath provisioning

Currently, many large-scale, resource-intensive applications and services are being developed that can be supported only with high-performance, highly distributed, heterogeneous infrastructures, including grids. This type of infrastructure is particularly effective for supporting applications and services that must quickly adjust to continuously changing conditions. Such processes require the flexibility of highly adaptive, dynamic, and deterministic resource provisioning. One such architecture is described here. To enhance the performance and flexibility of distributed environments, an experimental architecture for optical dynamic intelligent network (ODIN) services has been designed to enable core optical network capabilities to extend to edge processes, including applications. This architecture allows those processes to directly address arid control core network resources, for example, individual lightpaths on demand. This approach supports flexible and deterministic communications by integrating signaled requirements with adjustable network resources. An experimental prototype of ODIN has been designed, developed, and implemented on several optical network testbeds.     

支持区分服务的全光网络设计和实现

随着光层业务对传送服务质量的要求趋于多样化以及全光网络的智能化演进，光服务等级协定被提出。为了适应各种业务的不同传送需求，未来的光网络碰具有支持区分服务的能力。在多业务环境下。处理好满足业务的不同服务需求和优化网络资源之同的矛盾。需要传送网络流量工程各个功能子模块具有集成多种实现机制的能力，从而在光层实现区分服务的集成化。目前，支持区分服务的光网络流星工程方案还只是研究领域的成果，需要对现有控制平面的协议进行扩展和标准化，从而使光服务等级协定走向实用。     

Experimental demonstration of time-aware software defined networking for OpenFlow-based intra-datacenter optical interconnection networks

Nowadays, most service providers offer their services and support their applications through federated sets of data centers which need to be interconnected using high-capacity optical networks in intra-datacenter networks. Many datacenter applications in the environment require lower delay and higher availability with the end-to-end guaranteed quality of service. In this paper, we propose a novel time-aware software defined networking (TaSDN) architecture for OpenFlow-based intra-datacenter optical interconnection networks. Based on the proposed architecture, a time-aware service scheduling (TaSS) strategy is introduced to allocate the network and datacenter resources optimally, which considers the datacenter service scheduling with flexible service time and service bandwidth according to the various time sensitivity requirements. The TaSDN can arrange and accommodate the applications with required QoS considering the time factor, and enhance the data center responsiveness to quickly provide for intra-datacenter service demands. The overall feasibility of the proposed architecture is experimentally verified on our testbed with real OpenFlow-enabled tunable optical modules. The performance of TaSS strategy under heavy traffic load scenario is also evaluated based on TaSDN architecture in terms of blocking probability and resource occupation rate.     

The intelligent building telecommunications infrastructure

The telecommunications infrastructure required for an intelligent building to support the overall telecommunications utility is considered, focusing on the wiring system. Emerging national standards that embrace all components of the telecommunications infrastructure are described. Basic infrastructure concepts are explained, and component details are examined. Important links between the infrastructure components and the emerging national standards are shown     

Fair Class-Based Downlink Scheduling with Revenue Considerations in Next Generation Broadband Wireless Access Systems

The success of emerging Broadband Wireless Access Systems (BWASs) will depend, among other factors, on their ability to manage their shared wireless resources in the most efficient way. This is a complex task due to the heterogeneous nature, and hence, diverse Quality of Service (QoS) requirements of different applications that these systems support. Therefore, QoS provisioning is crucial for the success of such wireless access systems. In this paper, we propose a novel downlink packet scheduling scheme for QoS provisioning in BWASs. The proposed scheme employs practical economic models through the use of novel utility and opportunity cost functions to simultaneously satisfy the diverse QoS requirements of mobile users and maximize the revenues of network operators. Unlike existing schemes, the proposed scheme is general and can support multiple QoS classes with users having different QoS and traffic demands. To demonstrate its generality, we show how the utility function can be used to support three different types of traffic, namely best-effort traffic, traffic with minimum data rate requirements, and traffic with maximum packet delay requirements. Extensive performance analysis is carried out to show the effectiveness and strengths of the proposed packet scheduling scheme.     

Latency-driven distribution: infrastructure needs of participatory entertainment applications

This article evaluates network and server infrastructure requirements to support real-time flows associated with networked entertainment applications. These include the state information flow to update the status of the virtual environment and immersive communication flows such as voice, video, gesture, and haptics communication. The article demonstrates that scaling these applications to large geographical spreads of participants requires distribution of computation to meet the latency constraints of the applications. This latency-driven distribution of computation is essential even when there are no limitations on the availability of computational resources in one location. The article provides detailed results on distributed server architectures for two of these real-time flows, state information and immersive voice communication. It also identifies a generic set of requirements for the underlying network and server infrastructure to support these applications and propose a new design, called switched overlay networks, for this purpose.     

Application-Oriented Flow Control: Fundamentals, Algorithms and Fairness

This paper is concerned with flow control and resource allocation problems in computer networks in which real-time applications may have hard quality of service (QoS) requirements. Recent optimal flow control approaches are unable to deal with these problems since QoS utility functions generally do not satisfy the strict concavity condition in real-time applications. For elastic traffic, we show that bandwidth allocations using the existing optimal flow control strategy can be quite unfair. If we consider different QoS requirements among network users, it may be undesirable to allocate bandwidth simply according to the traditional max-min fairness or proportional fairness. Instead, a network should have the ability to allocate bandwidth resources to various users, addressing their real utility requirements. For these reasons, this paper proposes a new distributed flow control algorithm for multiservice networks, where the application's utility is only assumed to be continuously increasing over the available bandwidth. In this, we show that the algorithm converges, and that at convergence, the utility achieved by each application is well balanced in a proportionally (or max-min) fair manner     

Scalable routing strategies for ad hoc wireless networks

We consider a large population of mobile stations that are interconnected by a multihop wireless network. The applications of this wireless infrastructure range from ad hoc networking (e.g., collaborative, distributed computing) to disaster recovery (e.g., fire, flood, earthquake), law enforcement (e.g., crowd control, search-and-rescue), and military (automated battlefield). Key characteristics of this system are the large number of users, their mobility, and the need to operate without the support of a fixed (wired or wireless) infrastructure. The last feature sets this system apart from existing cellular systems and in fact makes its design much more challenging. In this environment, we investigate routing strategies that scale well to large populations and can handle mobility. In addition, we address the need to support multimedia communications, with low latency requirements for interactive traffic and quality-of-service (QoS) support for real-time streams (voice/video). In the wireless routing area, several schemes have already been proposed and implemented (e.g., hierarchical routing, on-demand routing, etc.). We introduce two new schemes-fisheye state routing (FSR) and hierarchical state routing (HSR)-which offer some competitive advantages over the existing schemes. We compare the performance of existing and proposed schemes via simulation     

Architectures,protocols and design for highly dynamic optical networks

In order for dynamic optical networks to be commercially viable, it will be essential to have signaling protocols that can establish connections quickly and minimize the use of transponders for wavelength conversion and regeneration. Rough estimates indicate that in dynamic optical networks, transponders will comprise a significant component (e.g., 60%–70%) of the cost of the optical networking infrastructure. A key performance metric in dynamic networks is the connection blocking probability, and therefore a dynamic optical network infrastructure must be designed to meet stringent blocking requirements (e.g., <10^?3 blocking probability). A key element in achieving that blocking objective is the ability to design the network to support shared transponder pools at selected network nodes, and have the capability to efficiently size those transponder pools to meet the required blocking probabilities. This paper addresses how to size shared transponder pools once the nodes that support those pools have been determined. The methodology can efficiently design for very small transponder-related blocking probability (e.g., <10^?4) by using simple, straightforward simulation and analysis techniques. We also describe a 3-Way-Handshake (3WHS) signaling protocol that can quickly set up connections (e.g., in <100 ms) and identify available resources (available wavelengths, transponders) so the connection setup process minimizes the use of transponders.     

Network resource management for enterprise wide multimedia services

The support of broadband multimedia applications over an enterprise network poses a broad range of networking challenges for efficient resource management, intelligent switching, and access control for distributed information. We propose two server-based scheduling algorithms, with low computational complexity, to guarantee multimedia information synchronization at the destination, with minimum presentation delays and buffer requirements. We outline a communication framework, highlighting the issues and challenges faced by today's enterprise networks to support multimedia services. We describe the proposed connection establishment and resource allocation schemes in resource-constrained enterprise networks. The objective is to manage the limited resources of the network for maximum utilization. Towards this end we present a dynamic capacity allocation scheme to support connections for multiple users. Specifically, we show that the channel capacity allocation problem can be formulated as a quadratic programming problem. This allocation scheme is implemented at each intermediate switch to dynamically determine the capacity allocation. The effects of interswitch rate mismatch and network delay offset scheduling have also been incorporated in the management framework. In addition, we introduce the concept of route selection based on the requested network quality of service     

虚拟网的资源管理

随着因特网的迅速发展,各种实时多媒体应用如远程教育、分布式仿真和网络游戏等得到了普通地应用,这些应用对网络系统有服务质量提出了较高的要求。如何通过网络资源的合理分配来满足不同应用之间的服务质量要求,是一个很重要的问题。本文利用虚拟网的概念来实现对网络资源的逻辑映射和统一管理,允许网络服务供应商和大公司客户实现灵活的可配置化管理,同时提高网络资源的复用效率。     

A fuzzy modeling based dynamic resource allocation strategy in service grid

One of the major challenges in service grid is to guarantee the promised quality of service (QoS) for all the admitted users or applications, while maximizing the resource utilization through dynamic resource sharing. An efficient resource allocation method should ensure the service QoS and balance the load among service grid nodes which are often highly dynamic, heterogeneous and linked by wide-area network. In this paper, a new dynamic resource allocation method is presented and analyzed based on fuzzy modeling to solve the adaptation between heterogeneous applications with multiple QoS requirements and grid resource. Simulations in service grid with heterogeneous QoS requirements reveal that the proposed dynamic resource allocation method can distribute most suitable resource among the different service quickly and sensitively as the service QoS demand varies under the constraint of achieving maximum utilization of grid resources.     

NFV技术在未来网络试验设施及运营商网络中的应用

未来网络试验设施作为我国在网络信息领域唯一的重大科技基础设施,其面向各类用户不同试验支持的需求使网络功能虚拟化(NFV)技术在其建设中得到重要的应用。阐述了NFV技术在试验设施从设备研发、网络组网到服务支持等方面的应用,并对其相关产品在当前运营商网络中的应用进行了分析和研究。     

Load balancing routing with bandwidth-delay guarantees

The current generation of network carriers competes intensely to satisfy the diverse wide-area connectivity requirements of customers. At the same time, the carriers inherently wish to maximize the usage efficiency of their network infrastructure. Much of the research in network resource management has been devoted to providing bandwidth guarantees and preventing network congestion. However, the rapid growth in number and diversity of real-time network applications has made it imperative to consider the impact of end-to-end delay of traffic requirements on network resource provisioning. We present an efficient network resource provisioning algorithm, called link criticality based routing (LCBR), which relies on the guiding theme that load balancing leads to higher resource utilization efficiency. LCBR applies a simple but very effective notion of link criticality to achieve networkwide load balance while simultaneously meeting the QoS requirements of bandwidth and end-to-end delay. In addition, LCBR can simultaneously provision both primary and backup routes to support fast recovery from node or link failures. This article reviews the state of the art in network resource provisioning with QoS guarantees, introduces the LCBR algorithm, and identifies future research challenges.     

Quality-of-service support for multimedia applications

A major project at Lancaster University is the development of network infrastructures capable of supporting the quality-of-service (QoS) requirements of a wide range of distributed multimedia applications. The project includes more than thirty researchers and covers not only the network support but also the enabling technologies. The authors believe that QoS development cannot be done in isolation from the applications to be supported, which must form an integral part of the project. They focus on three application areas: interactive teaching and learning, mobile systems, and virtual reality. They chose applications that stretch network support to its limit. To realize the full potential of these distributed multimedia applications, the underlying network must satisfy all these requirements concurrently. They give a brief overview of their activities to this end and discuss the need for QoS support