No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Market-Based Cooperative Resource Allocation for Overlay Networks
Abstract
In recent years, the notion of service overlay networks has been proposed as a promising solution for providing end-to-end QoS without changing the current Internet architecture. A major issue in deploying service overlay networks is determining how to allocate resources (such as link bandwidth) on a substrate network to overlay networks, while satisfying the end-to-end QoS requirements of applications running on each overlay network. This paper introduces the Market-based Cooperative Resource Allocation (MaCRA) architecture that achieves fair and efficient resource allocation in a decentralized manner. In MaCRA, resources on a substrate network are priced, and each overlay network provider creates an overlay network on a minimum cost basis to meet its application QoS requirements. MaCRA also allows each overlay network provider to trade their current resources with other overlay network providers when resources on a substrate network are not available or expensive. Simulation results demonstrate that MaCRA achieves fairness and efficiency in allocating resources for overlay networks when compared to existing mechanisms.
... Network device resources such as computation power, memory access time, storage size, and bandwidth are limited and expensive. Thus, resources must be efficiently managed to extract maximal benefit in various application scenarios [3], [2], [1], [4]. ...
Network resources are often limited, so how to use them efficiently is an issue that arises in many important scenarios. Many recent proposals rely on a central controller to carefully orchestrate resources across multiple network locations. The central controller gathers network information and relative levels of usage of different resources and calculates optimized task allocation arrangements to maximize some global benefit. Examples of architectures that use this framework include coordinated sampling (CSAMP [1]) and redundancy elimination (SmartRE [2]). However, a centralized solution creates practical problems as it is susceptible to overload, and the controller is a single point of failure. In this paper, we present a distributed solution called DECOR that achieves global optimization based on local information that closes to centralized approaches in terms of performance. In DECOR, the responsibility of resource monitoring and information gathering is spread among multiple nodes; thus, no single point is overloaded. Allocation of tasks is also done in a similar distributed fashion. DECOR can easily scale up to large networks, and the partial network failures do not affect DECOR's functioning in other parts of the network. DECOR can be applied to most of path-based applications. We describe in detail how to apply it to distributed SmartRE and implement it in the Click software router.
Simultaneous overlay networks compete for phys-ical network resources and disrupt each other. If they could establish cooperative relationships, the collective performance can be improved and they can coexist peacefully and comfortably. Taking inspiration from biology, in this paper we present a model of symbiotic overlay networks. Coexisting overlay networks dynamically evolve, interact with each other, and change their internal structures. Overlay networks in a symbiotic condition, i.e., mutualism, eventually establish the strong relationship and finally fuse into one. We first analyze characteristics of an overlay network which evolves based on three different models, i.e., the preferential attachment, random, and combination of them, by using mathematical analysis and simulation experiments. Next, we evaluate the effect of interconnecting two overlay networks from the viewpoint of the cost and the benefit.
In virtual circuit networks, the problem of routing a connection consists of selecting the best path among all the available paths between the source and destination of the connection. In this paper, we propose a path pre-selection routing scheme that not only takes the current state of each network link into account, but also considers the routing requests being issued by other nodes in the network. To do so, time is divided into periods. At each period, each node collects all its local connection requests, and broadcasts this list of requests to all nodes in the network. At the beginning of each period, all nodes apply the same path selection algorithm to choose the optimal paths for their requests from the previous period, taking into consideration the requests of all other nodes. Finding the optimal paths is formulated and solved as a linear integer optimization problem. Our algorithm outperforms a cost function routing algorithm in terms of setup blocking rates under different traffic loads and network configurations.
Consider a setting in which a group of nodes, situated in a large underlying network, wishes to reserve bandwidth on which to support communication. Virtual private networks (VPNs) are services that support such a construct; rather than building a new physical network on the group of nodes that must be connected, bandwidth in the underlying network is reserved for communication within the group, forming a virtual “sub-network.”Provisioning a virtual private network over a set off terminals gives rise to the following general network design problem. We have bounds on the cumulative amount of traffic each terminal can send and receive; we must choose a path for each pair of terminals, and a bandwidth allocation for each edge of the network, so that any traffic matrix consistent with the given upper bounds can be feasibly routed. Thus, we are seeking to design a network that can support a continuum of possible traffic scenarios.We provide optimal and approximate algorithms for several variants of this problem, depending on whether the traffic matrix is required to be symmetric, and on whether the designed network is required to be a tree (a natural constraint in a number of basic applications). We also establish a relation between this collection of network design problems and a variant of the facility location problem introduced by Karger and Minkoff; we extend their results by providing a stronger approximation algorithm for this latter problem.
Distributed Algorithmic Mechanism Design (DAMD) combines theoretical computer science's traditional focus on computational tractability with its more recent interest in incentive compatibility and distributed computing. The Internet's decentralized nature, in which distributed computation and autonomous agents prevail, makes DAMD a very natural approach for many Internet problems. This paper first outlines the basics of DAMD and then reviews previous DAMD results on multicast cost sharing and interdomain routing. The remainder of the paper describes several promising research directions and poses some specific open problems.
In recent years, service overlay networks have been proposed as a promising solution to provide end-to-end QoS without changing the current Internet architecture. In deploying service overlay networks, one major issue is how to efficiently allocate resources (e.g., link bandwidth) on a substrate network to each application with end-to-end QoS requirement. This paper proposes a novel decentralized architecture that achieves efficient resource allocation using a market mechanism. Resources on a substrate network are priced and each application provider aims to buy resources needed to satisfy its QoS requirement with minimum cost, which leads to globally efficient resource allocation. In order to further enhance efficiency in resource allocation, the proposed architecture enables trade of resources between applications. Trade of resources allows an application provider to buy resources from other application providers when resources on a substrate network are unavailable or expensive. Simulation results demonstrate the validity of the proposed architecture in improving efficiency of resource allocation.
This paper introduces the inter-domain Resource Exchange (iREX) architecture for the automated deployment of fault tolerant end to end (E2E) inter-domain (ID) quality of service (QoS) policy among resource user and resource provider Internet Service Providers (ISPs). iREX uses economics and fully distributed mechanisms to empower domains to self-manage the deployment of E2E ID QoS policy. In iREX, resource user domains select and reserve ID QoS resources to form E2E ID QoS policy while resource provider domains provide information about available ID QoS resources and support the deployment of policies. iREX promotes accountability by enabling the establishment of bilateral agreements between a resource user ISP and all resource provider ISPs, and promotes congestion- avoidance by enabling a distributed resource selection process that selects the least congested ID deployment path. With iREX, domains cooperate to deploy ID QoS policy while maintaining their autonomy at all times. Simulation results show that iREX blocks fewer reservations and causes less congestion than the existing method, and that iREX is able to recover from faults.
The routing infrastructure of the Internet has be- come resistant to fundamental changes and the use of overlay networks has been proposed to provide additional flexibility and control. One of the most prominent configurable components of an overlay network is its topology, which can be dynami- cally reconfigured to accommodate communication requirements that vary over time. In this paper, we study the problem of determining dynamic topology reconfiguration for service overlay networks with dynamic communication requirement, and the ideal goal is to find the optimal reconfiguration policies that can minimize the potential overall cost of using an overlay. We start by observing the properties of the optimal reconfiguration policies through studies on small systems and find structures in the optimal reconfiguration policies. Based on these observations, we propose heuristic methods for constructing different flavors of reconfiguration policies, i.e., never-change policy, always-change policy and cluster-based policies, to mimic and approximate the optimal ones. Our experiments show that our policy construction methods are applicable to large systems and generate policies with good performance. Our work does not only provide solutions to practical overlay topology design problems, but also provides theoretical evidence for the advantage of overlay network due to its configurability.
Recent proposals for network virtualization provide a promising way to overcome the Internet ossification. The key idea of network virtualization is to build a diversified Internet to support a variety of network services and architectures through a shared substrate. A major challenge in network virtualization is the assigning of substrate resources to virtual networks (VN) efficiently and on-demand. This paper focuses on two versions of the VN assignment problem: VN assignment without reconfigu- ration (VNA-I) and VN assignment with reconfiguration ( VNA- II). For the VNA-I problem, we develop a basic scheme as a building block for all other advanced algorithms. Subdividing heuristics and adaptive optimization strategies are then presented to further improve the performance. For the VNA-II problem, we develop a selective VN reconfiguration scheme that prioritizes the reconfiguration of the most critical VNs. Extensive simulation ex- periments demonstrate that the proposed algorithms can achieve good performance under a wide range of network conditions.
A Resilient Overlay Network (RON) is an architecture that allows distributed Internet applications to detect and recover from path outages and periods of degraded performance within several seconds, improving over today's wide-area routing protocols that take at least several minutes to recover. A RON is an application-layer overlay on top of the existing Internet routing substrate. The RON nodes monitor the functioning and quality of the Internet paths among themselves, and use this information to decide whether to route packets directly over the Internet or by way of other RON nodes, optimizing application-specific routing metrics. Results from two sets of measurements of a working RON deployed at sites scattered across the Internet demonstrate the benefits of our architecture. For instance, over a 64-hour sampling period in March 2001 across a twelve-node RON, there were 32 significant outages, each lasting over thirty minutes, over the 132 measured paths. RON's routing mechanism was able to detect, recover, and route around all of them, in less than twenty seconds on average, showing that its methods for fault detection and recovery work well at discovering alternate paths in the Internet. Furthermore, RON was able to improve the loss rate, latency, or throughput perceived by data transfers; for example, about 5% of the transfers doubled their TCP throughput and 5% of our transfers saw their loss probability reduced by 0.05. We found that forwarding packets via at most one intermediate RON node is sufficient to overcome faults and improve performance in most cases. These improvements, particularly in the area of fault detection and recovery, demonstrate the benefits of moving some of the control over routing into the hands of end-systems.
This paper describes the design, implementation, and ex- perimental evaluation of OverQoS, an overlay-based archi- tecture for enhancing the best-effort service of today's In- ternet. Using a Controlled loss virtual link (CLVL)abstrac- tion to bound the loss rate observed by a traffic aggregate, OverQoS can provide a variety of services including: (a) smoothing packet losses; (b) prioritizing packets within an aggregate; (c) statistical loss and bandwidth guarantees. We demonstrate the usefulness of OverQoS using two sam- ple applications. First, RealServer can use OverQoS to im- prove the signal quality of multimedia streams by protect- ing more important packets at the expense of less impor- tant ones. Second, Counterstrike, a popular multi-player game, can use OverQoS to avoid frame drops and prevent end-hosts from getting disconnected in the presence of loss rates as high as 10%. Using a wide-area overlay testbed of 19 hosts, we show that: (a) OverQoS can simultaneously provide statistical loss guarantees of 0.1% coupled with statistcal bandwidth guarantees ranging from 100 Kbps to2 Mbps across international links and broadband end-hosts; (b) OverQoS incurs a low bandwidth overhead (typically less than 5%) to achieve the target loss rate, and (c) the increase in the end-to-end delay is bounded by the round- trip-time along the overlay path.
Service Overlay Network (SON) was proposed to alleviate the difficulties encountered in providing end-to-end Quality of Service (QoS) guarantees. SON is able to provide QoS guarantees by purchasing bandwidth from individual network domains and building a logical end-to-end data delivery infrastructure on top of the existing Internet. We focus on SON topology design problems under a generalized cost model. Earlier research in this topic considered two distinct cost models - fixed (leased) cost model and variable (usage-based) cost model. However in most applications, the costs of both nodes and links have a fixed component as well as a variable component that often depends on usage. Our generalized cost model takes this fact into account and our topology design algorithm uses this cost model to And the optimal topology. Since the SON topology design problem is NP-complete, we provide approximation algorithm with guaranteed performance bound. We validate the effectiveness of our algorithm through extensive simulation.
The Internet still lacks adequate support for QoS applications with real-time requirements. In great part, this is due to the fact that provisioning of end-to-end QoS to traffic that traverses multiple autonomous systems (ASs) requires a level of cooperation between ASs that is difficult to achieve in the current architecture. Recently, service overlay networks have been considered as an approach to QoS deployment that avoids these difficulties. In this study, we address the problem of the topological synthesis of a service overlay network, where endsystems and nodes of the overlay network (provider nodes) are connected through ISPs that supports bandwidth reservations. We express the topology design problem as an optimization problem. Even though the design problem is related to the (in general NP-hard) quadratic assignment problem, we are able to show that relatively simple heuristic algorithms can deliver results that are sometimes close to the optimal solution.
We advocate the notion of service overlay network (SON) as an effective means to address some of the issues, in particular end-to-end QoS, plaguing the current Internet, and to facilitate the creation and deployment of value-added Internet services such as VoIP, video-on-demand, and other emerging QoS-sensitive services. A SON purchases bandwidth with certain QoS guarantees from individual network domains via a bilateral service level agreement (SLA) to build a logical end-to-end service delivery infrastructure on top of existing data transport networks. Via a service contract, users directly pay the SON provider for using the value-added services provided by the SON. We study the bandwidth provisioning problem for a service overlay network which is critical to the cost recovery in deploying and operating value-added services over the SON. We mathematically formulate the bandwidth provisioning problem, taking into account various factors such as SLA, service QoS, traffic demand distributions, and bandwidth costs. Analytical models and approximate solutions are developed for both static and dynamic bandwidth provisioning. Numerical studies are also performed to illustrate the properties of the proposed solutions and demonstrate the effect of traffic demand distributions and bandwidth costs on the bandwidth provisioning of a SON.
We advocate the notion of service overlay network (SON) as an effective means to address some of the issues, in particular, end-to-end quality of service (QoS), plaguing the current Internet, and to facilitate the creation and deployment of value-added Internet services such as VoIP, Video-on-Demand, and other emerging QoS-sensitive services. The SON purchases bandwidth with certain QoS guarantees from the individual network domains via bilateral service level agreement (SLA) to build a logical end-to-end service delivery infrastructure on top of the existing data transport networks. Via a service contract, users directly pay the SON for using the value-added services provided by the SON. In this paper, we study the bandwidth provisioning problem for a SON which buys bandwidth from the underlying network domains to provide end-to-end value-added QoS sensitive services such as VoIP and Video-on-Demand. A key problem in the SON deployment is the problem of bandwidth provisioning, which is critical to cost recovery in deploying and operating the value-added services over the SON. The paper is devoted to the study of this problem. We formulate the bandwidth provisioning problem mathematically, taking various factors such as SLA, service QoS, traffic demand distributions, and bandwidth costs. Analytical models and approximate solutions are developed for both static and dynamic bandwidth provisioning. Numerical studies are also performed to illustrate the properties of the proposed solutions and demonstrate the effect of traffic demand distributions and bandwidth costs on SON bandwidth provisioning.
Recently, many overlay applications have emerged in the Internet. Currently, each of these applications requires their proprietary functionality support. A general unified framework may be a desirable alternative to application-specific overlays. We introduce the concept of overlay brokers (OBs). We assume that each autonomous system in the Internet has one or more OBs. These OBs cooperate with each other to form an overlay service network (OSN) and provide overlay service support for overlay applications, such as resource allocation and negotiation, overlay routing, topology discovery, and other functionalities. The scope of our effort is the support of quality-of-service (QoS) in overlay networks. Our primary focus is on the design of QoS-aware routing protocols for overlay networks (QRONs). The goal of QRON is to find a QoS-satisfied overlay path, while trying to balance the overlay traffic among the OBs and the overlay links in the OSN. A subset of OBs, connected by the overlay paths, can form an application specific overlay network for an overlay application. The proposed QRON algorithm adopts a hierarchical methodology that enhances its scalability. We analyze two different types of path selection algorithms. We have simulated the protocols based on the transit-stub topologies produced by GT-ITM. Simulation results show that the proposed algorithms perform well in providing a QoS-aware overlay routing service.