DiffServ Networks Research Articles

A flexible and distributed architecture for adaptive end-to-end QoS provisioning in next-generation networks

A novel distributed end-to-end quality-of-service (QoS) provisioning architecture based on the concept of decoupling the end-to-end QoS provisioning from the service provisioning at routers in the differentiated service (DiffServ) network is proposed. The main objective of this architecture is to enhance the QoS granularity and flexibility offered in the DiffServ network model and improve both the network resource utilization and user benefits. The proposed architecture consists of a new endpoint admission control referred to as explicit endpoint admission control at the user side, the service vector which allows a data flow to choose different services at different routers along its data path, and a packet marking architecture and algorithm at the router side. The achievable performance of the proposed approach is studied, and the corresponding results demonstrate that the proposed mechanism can have better service differentiation capability and lower request dropping probability than the integrated service over DiffServ schemes. Furthermore, it is shown that it preserves a friendly networking environment for conventional transmission control protocol flows and maintains the simplicity feature of the DiffServ network model.

Fair intelligent admission control over resource-feedback DiffServ network

The basic DiffServ model lacks mechanisms to prevent itself from being overloaded and to inform its internal capability to the external world. This paper addresses the problem by presenting a Fair Intelligent Admission Control (FIAC) over an enhanced-DiffServ architecture. The central idea is to make admission decision based on both informed internal network QoS states and the external traffic QoS requirements at the edge node. This model has several advantages: (1) it is backward compatible with DiffServ, (2) it adapts to traffic load and network QoS state changes, and (3) it provides interactive communication between the external QoS requirements and the internal DiffServ network capability. In this paper, we use simulation to evaluate the performance of DiffServ with or without FIAC. The performance demonstrates that the new scheme is able to admit traffic fairly and achieve edge-to-edge QoS under heavy traffic conditions and network state changes.

Statistical admission control for real-time services under earliest deadline first scheduling

The design of call admission control (CAC) mechanisms has been a critical issue in providing quality-of-service (QoS) guarantees for heterogeneous traffic flows over integrated service (IntServ) and/or differentiated service (DiffServ) networks. Earliest deadline first (EDF) is an ideal scheduler for real-time services because of its optimal admissible region and delay bound properties. The major difficulty in developing an effective and efficient CAC algorithm for statistical services is the analysis of per-class deadline violation (loss) probabilities with respect to the delay bounds. In this paper, we provide an analytical approach to evaluate the aggregate and per-class deadline violation probabilities of an EDF scheduler. Based on these theoretical foundations, we derive the admission control conditions and then propose a CAC algorithm for statistic services under EDF scheduling. In addition, we show that the QoS metrics that an EDF scheduler actually guarantees have an asymptotic ordering property, which provides an important insight into the design and control of EDF networks. The effectiveness and performance of our proposed algorithm have been validated by trace-driven simulation experiments using MPEG and H.263 encoded video sources.

Active Queue Management Algorithms in DiffServ Networks

Current assured forwarding (AF) service in differentiated services (DiffServ) networks can provide stable guarantees in throughput, but is lacking of efficient schemes in ensuring queuing delay and loss ratio. By analyzing the steady state operating point of RIO, this paper proposes two active queue management algorithms with adaptive control policy, namely ARIO-D and ARIO-L. These two algorithms can provide differentiated performance in, respectively, queuing delay and loss ratio, in addition to throughput guarantee. By deploying ARIO-D and ARIO-L, AF service can provide quantitative guarantees for multimedia traffic with multiple QoS

SLA-based QoS pricing in DiffServ networks

The availability of high-speed transmission media and networking equipment in contemporary networks, as well as the evolution of quality-demanding applications has focused research interest on the provision of advanced qualitative services in addition to the traditional best-effort model of the Internet. A number of alternatives for service differentiation and QoS provisioning have been proposed and standardized, but in the case of backbone, transport networks the DiffServ architecture has prevailed, due to its scalability and deployment feasibility. The provisioning of services according to the DiffServ framework has in turn raised the requirements for interdependent, controlled resource allocation and service pricing, with particular needs for pricing mechanisms that preserve the potential and flexibility of DiffServ. At the same time, such mechanisms should reflect resource usage, allocate resources efficiently, reimburse costs or maximize service provision profits and lead customers to requesting services that will maximize their revenue. In this work, after reviewing related research, the principles that a pricing scheme for DiffServ-based services should follow are presented, stressing the differences form traditional Internet pricing. Based on these principles, an analytical approach to pricing a particular class of DiffServ-based services and a methodology for applying this approach in a real network are proposed and evaluated.

A measurement-based approach for dynamic QoS adaptation in DiffServ networks

This paper presents a practical approach to managing multimedia traffic in DiffServ network, using network monitoring feedback and control. We exploit the flexibility of multimedia traffic and process network level parameters to adapt the traffic according to the current state of the network. The latter is determined based on reports sent by bandwidth monitors installed on each node of a DiffServ Domain. The bandwidth monitors interact with a policy server which, depending on the network state, decides the policy(ies) that should be enforced by the DiffServ network. The implementation of the selected policies typically leads to accepting, remarking, or dropping the multimedia traffic entering the network. Multimedia streams may be assigned different levels of QoS, as interpreted by the marker at the DiffServ edge router and marked according to network state. To achieve such dynamic QoS adaptation for multimedia applications, we have implemented and evaluated a policy-based management system. Performance evaluation shows that multimedia applications adapt better to network conditions using our approach.

Packet size based dynamic scheduling for assured services in Diffserv network

AbstractDiffserv (Differentiated Service) [1] is a promising technique to realize scalable QoS control in the Internet. Therefore, extensive efforts have been made regarding the practical use of this technique. Diffserv has already been introduced in some networks. Control functions for Diffserv are installed in the edge router and core router. This enables a scalable QoS control as mentioned above. Herein we propose a new active queue management scheme for AF services guaranteeing minimum bandwidth provided that RIO and WFQ are used as the basic QoS control mechanisms. The proposed method dynamically controls queue threshold values of RIO and bandwidth allocation of WFQ taking into account the packet sojourn time and packet size of each head of line packet. Simulation results show that the proposed scheme reduces the packet delay time for higher‐priority QoS classes when packets with longer size are frequently transmitted. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 1, 88(1): 12–20, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecja.20151

Inter-domain QoS routing on Diffserv networks: a region-based approach

Quality of Service (QoS) routing is inherently a difficult problem. Inter-domain QoS routing is even harder, because it involves entities residing in distinct administrative domains. There are two problems that need to be solved in inter-domain QoS routing: topology distribution in a scalable fashion and finding paths that satisfy QoS constraints and provide connectivity. In this paper we present region-based, link-state, source-specified, inter-domain QoS routing architecture that addresses these questions. Our architecture is scalable and does not suffer from the problems caused by hierarchical routing. Analysis results show that the average region size and the average shortest path length (SPL) are inversely proportional and scalability of the approach increases as the region size decreases. Gain from the scalability is far more than the loss from the average SPL, especially with larger topologies.

Assured end-to-end QoS through adaptive marking in multi-domain differentiated services networks

The issue of resource management in multi-domain Differentiated Services (DiffServ) networks has attracted a lot of attention from researchers who have proposed various provisioning, adaptive marking and admission control schemes. In this paper, we propose a Reinforcement Learning-based Adaptive Marking (RLAM) approach for providing assured end-to-end quality of service (QoS) in the form of end-to-end delay and throughput assurances, while minimizing packet transmission cost since ‘expensive’ Per Hop Behaviors like Expedited Forwarding (EF) are used only when necessary. The proposed scheme tries to satisfy per flow end-to-end QoS through control action,s which act on flow aggregates in the core of the network. Using an ns2 simulation of a multi-domain DiffServ network with multimedia traffic, the RLAM scheme is shown to be effective in significantly lowering packet transmission costs without sacrificing end-to-end QoS, when compared to the commonly used static marking scheme.

Performance of Hierarchical Aggregation in Differentiated Services Networks

We address the use of hierarchical aggregation in DiffServ networks. We propose two analytical models to study the tradeoffs between signaling load and resource utilization. In the case of the signaling load, we introduce a novel performance metric that captures, simultaneously, the state information stored and the rate of signaling messages processed at routers. In the first analytical model, based on multidimensional birth---death processes, the offered load is detailed at the flow level, which allows accurate assessment of the signaling load. The second analytical model accommodates time-varying offered loads, which allows studying the tradeoffs between the time-scale of the aggregate demand and the time-scale of signaling. Our results, which also include analysis using measured traces, show that hierarchical aggregation can introduce very high signaling gains with a small penalty in terms of resource utilization, allowing significant savings in terms of network cost.

SLA-based QoS pricing in DiffServ networks

The availability of high-speed transmission media and networking equipment in contemporary networks, as well as the evolution of quality-demanding applications has focused research interest on the provision of advanced qualitative services in addition to the traditional best-effort model of the Internet. A number of alternatives for service differentiation and QoS provisioning have been proposed and standardized, but in the case of backbone, transport networks the DiffServ architecture has prevailed, due to its scalability and deployment feasibility. The provisioning of services according to the DiffServ framework has in turn raised the requirements for interdependent, controlled resource allocation and service pricing, with particular needs for pricing mechanisms that preserve the potential and flexibility of DiffServ. At the same time, such mechanisms should reflect resource usage, allocate resources efficiently, reimburse costs or maximize service provision profits and lead customers to requesting services that will maximize their revenue. In this work, after reviewing related research, the principles that a pricing scheme for DiffServ-based services should follow are presented, stressing the differences form traditional Internet pricing. Based on these principles, an analytical approach to pricing a particular class of DiffServ-based services and a methodology for applying this approach in a real network are proposed and evaluated.

Pricing and admission control for QoS-enabled Internet

Over the past ten years, many pricing schemes have been proposed for a QoS-enabled network. Most of the proposed QoS-pricing schemes focus on congestion-sensitive pricing and optimal pricing solutions. Integrating pricing and admission control has not been studied in details. In this paper, we pay more attention to the interrelation between pricing and admission control in QoS-enabled networks and propose a tariff-based architecture framework that flexibly integrates pricing and admission control for multi-domain DiffServ networks. We study the pricing and user behaviors in detail and design a market-regulated pricing and admission control scheme in our framework. We model the system as a market so that the price of a service class reflects the resource availability inside the network and is regulated by the market itself. We also evaluate our pricing strategy and admission control scheme through simulations.

Scalable Architecture and Evaluation for Multiparty Conferencing Over Satellite Links

This paper describes the architecture of a proposed multiparty conferencing system for satellites. Different conferencing models are discussed and compared. A session initiation protocol (SIP)-based conference signaling model and an extension to protocol-independent multicast sparse mode that supports quality-of-service (QoS) in DiffServ networks are proposed, as particularly suitable for multiparty conferencing applications over satellite links. The paper also presents key issues and potential solutions of scalable QoS multicast services for multiparty conferences over satellite. End-to-end QoS parameters for voice and video are measured and analyzed on a prototype.

Capacity Planning of DiffServ Networks with Best-Effort and Expedited Forwarding Traffic

For networks providing a specific level of service guarantees, capacity planning is an imperative part of network management. Accurate dimensioning is especially important in DIffServ networks, where no per-flow signaling or control exists. In this paper, we address the problem of capacity planning for DIffServ networks with only expedited forwarding (EF) and best effort (BE) traffic classes, The problem is formulated as an optimization problem, where the total link cost is minimized, subject to the performance constraints of both EF and Be classes. The variables to be determined are the routing of EF traffic, and the discrete capacities of the links. We show that Lagrangean relaxation and subgradient optimization methods can be used to solve the problem. Computational results show that the solution quality is verifiably good while the running time remains reasonable on practical-sized networks. This represents the first work for capacity planning of multi-class IP networks with non-linear performance constraints.

Scalable Quasi-Dynamic-Provisioning-Based Admission Control Mechanism in Differentiated Service Networks

The architecture in a differentiated services (diffServ) network is based on a simple model that applies a per-class service in the core node of the network. However, because the network behavior is simple, the network structure and provisioning is complicated. If a service provider wants dynamic provisioning or a better bandwidth guarantee, the differentiated services network must use a signaling protocol with QoS parameters or an admission control method. Unfortunately, these methods increase the complexity. To overcome the problems with complexity, we investigated scalable dynamic provisioning for admission control in DiffServ networks. We propose a new scalable qDPM 2 mechanism based on a centralized bandwidth broker and distributed measurement-based admission control and movable boundary bandwidth management to support heterogeneous QoS requirements in DiffServ networks.

A Markovian Model for TCP Analysis in a Differentiated Services Network

In a Differentiated Services network, the use of TCP by an application impacts the service it gets from the network. TCP congestion control algorithms are designed to provide a fair sharing of resources in a best effort network as the current Internet. TCP is not conscious of the new services proposed by DiffServ, namely the different priorities packets are injected with into the network. Many schemes have been proposed to support TCP traffic in a DiffServ network. These schemes have been often validated with simulations. In this paper we propose an analytical model to study the performance of TCP in a DiffServ network under the different proposed schemes. The model is based on a Markovian fluid approach. We present first a general version of the model, then we specify it to the different proposed schemes. For each scheme, we compute the throughput achieved by a TCP connection. We compare then the service differentiation provided by the proposed schemesunder different subscription levels, different reservations, and different round-trip times. Our model forms a good tool for the evaluation of new solutions to support TCP traffic in a DiffServ network.

Enhancing end-to-end QoS granularity in DiffServ networks via service vector and explicit endpoint admission control

The DiffServ model is considered as a scalable and viable solution to provide QoS in the Internet. However, the current DiffServ model can only provide a coarse end-to-end QoS granularity to individual data flows. This is undesirable for many network services such as the virtual private network (VPN), since it may result in low resource utilisation to guarantee the required QoS. In the paper the concept of service vector is proposed to provide finer QoS granularity in DiffServ networks through the framework of a distributed admission control scheme referred to as explicit endpoint admission control (EEAC). Specifically, by applying the proposed service vector and EEAC scheme, a data flow is allowed to choose different services at different nodes to achieve fine QoS granularity and architectural scalability. The models and solutions that represent the users' behaviour in selecting the appropriate network services are also discussed. The performance evaluation of the proposed approach demonstrates that the service vector and EEAC scheme can result in lower request drop ratio, satisfactory end-to-end performance, fine end-to-end QoS granularity, and lower cost for each user.

Supporting real-time IP traffic with enhanced service classes in DiffServ networks

The differentiated services (DiffServ) architecture offers a scalable alternative to provide quality of service (QoS) guarantees for performance-sensitive applications in the Internet. Within the DiffServ framework, however, an efficient traffic scheduling mechanism is a key component to ensure such QoS guarantees. The authors propose a packet scheduling algorithm called enhanced weighted fair queueing (EWFQ), which enables fair bandwidth sharing while supporting tight bounds on end-to-end delay for real-time traffic such as voice over IP (VoIP) in DiffServ networks. EWFQ allows service classes to be created and proportional weights to be assigned to such classes efficiently according to their resource requirements. Results from simulation studies indicate that the mechanism proposed ensures both the required bandwidth fairness and end-to-end network delay bounds according to the specified weight ratios under various traffic and network conditions. The scheme also has lower implementation complexity, along with scalability to accommodate growing traffic flows at the core routers of the high-speed Internet backbone.

Improving aggregate flow control in differentiated services networks

The differentiated services (DiffServ) framework provides a simple and scalable methodology for addressing quality of service issues in IP networks. Service differentiation can be achieved through the assured forwarding (AF) per hop behavior packet-handling scheme. One of the more challenging research issues in DiffServ networks is the fair distribution of bandwidth among aggregates sharing the same AF class. Several studies have shown that the number of microflows in aggregates, the round trip time, the mean packet size and TCP/UDP interactions are key factors in the throughput obtained by aggregates using this architecture. In this paper, we examine the behavior of a recent technique suggested to improve fairness in a DiffServ network: the Aggregate Flow Control mechanism. We also propose two alternatives in the control overlay of this scheme and compare them with the original approach. Simulation results indicate that our proposed modifications improve throughput assurance and fairness requirements.

Dynamic resource configuration in DiffServ networks: control plane mechanisms and performance evaluation of a traffic control API

Voice, video and multimedia applications are sensitive to the QoS provided by the underlying IP network. The DiffServ architecture offers a set of QoS mechanisms for IP networks. The “binding” of the applications QoS needs with the QoS features offered by the DiffServ networks is still an open problem. The simplest approach is to have a static configuration of QoS and therefore no direct interaction in the control plane between applications and QoS. We consider the advanced scenario, where the QoS mechanisms can be dynamically configured to follow the applications’ need. For this scenario, a set of control plane interfaces needed for a whole end-to-end QoS architecture is defined. At the lower level, an internal interface (“Application Programming Interface”––API) in the QoS router is considered. This interface provides access to the DiffServ QoS mechanism available in a router and is used by the control logic running in the router itself. Then a QoS signaling protocol is considered, that allows external QoS clients to dynamically access the QoS services provided by the network. Finally the interaction of a session level signaling protocol (i.e., the SIP for IP telephony) with the QoS protocol is defined. A testbed implementation of the proposed architecture and a set of performance tests on the internal QoS API are reported.