Programmable Routers Research Articles

A generalized service deployment framework for OSPF networks using multi‐topology routing: Modeling and methodology

SummaryThe exponential growth of various applications requires deploying an ever‐growing number of network services. A generalized service deployment framework for Open Shortest Path First (OSPF) networks is proposed in this paper. The framework includes placing programmable routers, distributing different types of services on these routers, and leading traffic flow through them according to the predetermined sequence order requirement. However, it is not possible to direct all the traffic flows through the required service nodes along the shortest path with a single and suitable set of link weights. To address the issue, multiple topology routing (MTR) technique is incorporated to have various logical topologies with multiple sets of link weights. Correspondingly, the problem of jointly optimizing Placement of programmable routers, Distribution of different types of services among these routers, and Link Weights setting based on MTR (shortened to PD‐LW‐MTR) and its mixed integer linear programming formulation are presented in this paper. A novel decomposition algorithm is also proposed to address this problem efficiently. Experiment results validate the correctness and feasibility of our algorithm. It is also shown that the optimization algorithm can obtain near‐optimal solution and just only a few logical topologies over multiple sets of link weights are necessary for traffic flows to guarantee service order requirements.

PROGRES: A Programmable Green Router With Controlled Service Rate

Routers are the most common transmission devices over the Internet. Architecture of the routers can be improved to fully utilize the chances to save energy related to the CMOS based transmission equipment. One prime approach for controlling and managing the power dissipation of the CMOS devices is Dynamic Voltage and Frequency Scaling (DVFS). This paper implements the DVFS in a CMOS device programmed as a router for studying multiple factors impeding energy saving by using Dynamic Clock Generation (DCG) and Dynamic Clock Selection (DCS). The authors emphasize on the performance evaluation of such frequency scaled CMOS devices, in which the service rate vary according to the state of an underlying finite-state, continuous-time Markov model. Furthermore, this paper establishes the relationships of critical factors (in/)directly effecting power consumption of these devices. Aforementioned factors and their dependencies such as the arrival and service rates, queuing lengths and thresholds in such systems facilitate the changing dynamics of functions and services for appropriate clock state of the device. Thus, the resulting model is quite complex to be proved analytically or numerically. We investigate and implement PROGRES to affirm the numerical models' accuracy for capturing the service rate transitions of the proposed control policies with detailed performance analysis and discussion. Furthermore, PROGRES also scales the arrival rates and queue lengths, while ensuring the controlled transition rates of the modulating Markov process in the programmable router. Two control models against single service rate for benchmarking. These techniques and numerical experiments allow us to identify and evaluate the trade-offs between energy consumption and system performance in the futuristic programmable networking devices with resulting approximations are promising at network, node and functional levels.

Routing or Computing? The Paradigm Shift Towards Intelligent Computer Network Packet Transmission Based on Deep Learning

Recent years, Software Defined Routers (SDRs) (programmable routers) have emerged as a viable solution to provide a cost-effective packet processing platform with easy extensibility and programmability. Multi-core platforms significantly promote SDRs’ parallel computing capacities, enabling them to adopt artificial intelligent techniques, i.e., deep learning, to manage routing paths. In this paper, we explore new opportunities in packet processing with deep learning to inexpensively shift the computing needs from rule-based route computation to deep learning based route estimation for high-throughput packet processing. Even though deep learning techniques have been extensively exploited in various computing areas, researchers have, to date, not been able to effectively utilize deep learning based route computation for high-speed core networks. We envision a supervised deep learning system to construct the routing tables and show how the proposed method can be integrated with programmable routers using both Central Processing Units (CPUs) and Graphics Processing Units (GPUs). We demonstrate how our uniquely characterized input and output traffic patterns can enhance the route computation of the deep learning based SDRs through both analysis and extensive computer simulations. In particular, the simulation results demonstrate that our proposal outperforms the benchmark method in terms of delay, throughput, and signaling overhead.

A survey on future internet architecture

Internet has become as a social infrastructure.The current Internet architecture based on TCP/IP is faced with many challenges.This fact makes the clean-slate design of future Internet architecture be a hot research topic.This paper analyzes the essential problems that are inherent in the current Internet architecture,including scalability,mobility,security and controllability,and comprehensively introduces the existing research works on the future Internet architecture such as the scalability-oriented architecture,the mobility-oriented architecture etc.Considering experiment-based validation is an important means for the research of future Internet,this paper further presents the development of programmable virtual routers and the corresponding testbeds.Finally,open issues on the future Internet architecture design are discussed.

GrainFlow: Enable testing for future internet architectures by per-bit customization

Programmable routers are key components for building future Internet test beds. Extensive research, such as the OpenFlow project, has been carried out to enhance programmability in the control plane of routers. However, the data plane is inflexible to process arbitrarily-defined packets. Designing a general purpose data plane platform to support distinct networks has been a challenge that must be resolved. In this article, we propose GrainFlow, a fully programmable hardware-based packet processing platform. GrainFlow brings flexibility to data plane by arbitrarily defining operations on every bit of packet, and adapts to variation of packet header length by employing a configurable circular pipeline. We implement GrainFlow on a Field Programmable Gate Array (FPGA) board. Finally, we evaluate the resource utilization and network performance of the GrainFlow architectures presented.

Optical switch emulation in programmable software router testbed

A programmable optical router is a key enabler for dynamic service provisioning in Future Internet scenarios. It is equipped with optical switching hardware to forward information at hundreds of Gigabits/s rates and above, controlled and managed through modular and flexible procedures according to emerging standards. The possibility to test such costly optical architectures in terms of logical and physical performance, without implementing complex and expensive testbeds, is crucial to speed-up the development process of high-performance routers. To this purpose, this paper introduces the software-based emulation testbed of a programmable optical router, which is here developed and applied to test optical switching fabrics. Accurate characterization of the optical devices and physical layer aspects is implemented with the Click software router environment. Power loss and optical signal-to-noise-ratio evaluation are provided through accurate software representation of the physical characteristics of the optical devices employed. The scalability of the proposed emulation testbed is also assessed on standard PC hardware. All the obtained results prove the effectiveness of the proposed tool to emulate an optical router at different levels of granularity.

A Survey on Future Internet Architecture

Internet has become as a social infrastructure.The current Internet architecture based on TCP/IP is faced with many challenges.This fact makes the clean-slate design of future Internet architecture be a hot research topic.This paper analyzes the essential problems that are inherent in the current Internet architecture,including scalability,mobility,security and controllability,and comprehensively introduces the existing research works on the future Internet architecture such as the scalability-oriented architecture,the mobility-oriented architecture etc.Considering experiment-based validation is an important means for the research of future Internet,this paper further presents the development of programmable virtual routers and the corresponding testbeds.Finally,open issues on the future Internet architecture design are discussed.

An enabling platform for autonomic management of the future internet

This article shows an autonomic management solution based on the recently defined programmable node architecture NetServ. The article starts with a general description of the classical network management requirements and their adaptation to the expected network evolution. After a description of the major issues characterizing the management of the expected Future Internet, the main autonomic management paradigms, and some recently introduced autonomic service platforms, we show and demonstrate the effectiveness of the NetServ architecture. Born as a means to deploy and execute networked services at runtime over programmable routers, NetServ has proved to be a suitable environment for hosting an autonomic management architecture.

PEARL: a programmable virtual router platform

Programmable routers supporting virtualization are a key building block for bridging the gap between new Internet protocols and their deployment in real operational networks. This article presents the design and implementation of PEARL, a programmable virtual router platform with relatively high performance. It offers high flexibility by allowing users to control the configuration of both hardware and software data paths. The platform makes use of fast lookup in hardware and software exceptions in commodity multicore CPUs to achieve highspeed packet processing. Multiple isolated packet streams and virtualization techniques ensure isolation among virtual router instances.

Bypass Flow-Splitting Forwarding in FISH Networks

According to the fundamental interconnecting structure homogeneous network model and its routing oscillation phenomenon in Internet service provider digital ecosystems, a bypass flow-splitting forwarding (BFF) algorithm is proposed. The BFF algorithm utilizes the mirror process machine (MPM) to analyze the flow characteristics. Using cross-layer ideas, the MPM computes new forwarding granularity. Using the collaboration between MPM and programmable routers (PRs), BFF utilizes the type-of-service field to forward packets and makes full use of available bandwidth of multipath. Using real-world topology and traffic traces in NS2, the simulations prove that the BFF obtains a better forwarding performance with higher link utilization and lower delay. BFF can be easily deployed into the PRs of next-generation Internet.

Standard-based approach to programmable hybrid networks

Programmable hybrid networks are considered in this article as a challenge to deploy flexible solutions for next-generation networks. A modular programmable router architecture, which takes advantage of emerging standards, is proposed to support dynamic management and configuration capabilities of network resources and services. The new node design concepts allow vendor-independent development of hybrid router functionalities and possible investigation of new standard interfaces. Validation of the proposed approach is proven in an open source software router context, emulating a hybrid circuit/packet service based on optical switching technology. The software emulator obtained allows programmable hybrid router functionalities to be tested at different levels of abstraction and with different modeling granularities, and is able to represent both logical and physical aspects.

EXPLOITING A GENERIC APPROACH TO CONSTRUCT COMPONENT-BASED SYSTEMS SOFTWARE IN LINUX ENVIRONMENTS

Component-based software engineering has recently emerged as a promising solution to the development of system-level software. Unfortunately, current approaches are limited to specific platforms and domains. This lack of generality is particularly problematic as it prevents knowledge sharing and generally drives development costs up. In the past, we have developed a generic approach to component-based software engineering for system-level software called OpenCom. In this paper, we present OpenComL an instantiation of OpenCom to Linux environments and show how it can be profiled to meet a range of system-level software in Linux environments. For this, we demonstrate its application to constructing a programmable router platform and a middleware for parallel environments.

Towards a cost-effective networking testbed

The Internet is suffering from ossification. There has been substantial research on improving current protocols, but the vendors are reluctant to deploy new ones. We believe that this is in part due to the difficulty of evaluating protocols under realistic conditions. Recent wide-area testbeds can help alleviate this problem, but they require substantial resources (equipment, bandwidth) from each participant, and they have difficulty in providing repeatability and full control over the experiments. Existing in-house networking testbeds are capable of running controlled, repeatable experiments, but are typically small-scale (due to various overheads), limited in features, or expensive. The premise of our work is that it is possible to leverage the recent increases in computational power to improve the researchers' ability to experiment with new protocols in lab settings. We propose a cost-effective testbed, called MX, which emulates many programmable routers running over a realistic topology on multi-core commodity servers. We leverage open source implementations of programmable routers, such as Click, and modify them to allow coexistence of multiple instances in the same kernel in an effort to reduce packet forwarding overheads. Our initial results show that we outperform similar cost-effective solutions by a factor of 2. Next, we demonstrate that grouping and placing routers on to cores which share the L2 cache yields high performance.

New frontiers in internet network management

Network management represents an architectural gap in today's Internet [1]. Many problems with computer networks today, such as faults, misconfiguration, performance degradation, etc., are due to insufficient support for network management, and the problem takes on additional dimensions with the emerging programmable router paradigm. The Internet Network Management Workshop is working to build a community of researchers interested in solving the challenges of network management via a combination of bottoms-up analysis of data from existing networks and a top-down design of new architectures and approaches driven by that data. This editorial sets out some of the research challenges we see facing network management, and calls for participation in working to solve them.

A practical bytecode interpreter for programmable routers on IXP network processors

WASP is a programmable router platform that allows end-hosts to store ephemeral state in routers along the path of IP flows and to execute packet-attached bytecode that processes this data. We exploit lessons from past active network research and our knowledge of network processors to design a minimal interpreter that favours language restrictions over run-time checks. WASP provides safety with limited performance penalty through predictable execution time and bounded usage of memory and network resources. WASP is expressive enough to enable several applications including statistics collection and service discovery. It can also detect common trunk of two Internet paths and exchange local measurements about these paths. We propose a robust implementation on the IXP2400 network processor, and evaluate its performance through short benchmark programs against native functions hard-coded in the router. We achieve latencies below [email protected], i.e. less than the reference IPv4 forwarding latency, and throughputs approaching 800kpps per core, which competes with, and sometimes even outperforms, native programs. We further exploit our results to give hints on further improving resource usage and guidelines on management of ephemeral stores in high-speed networks.

Analysis and Implementation of an Open Programmable Router Based on Forwarding and Control Element Separation

A router architecture based upon ForCES (Forwarding and Control Element Separation), which is being standardized by IETF ForCES working group, gains its competitive advantage over traditional router architectures in flexibility, programmability, and cost-effectiveness. In this paper, design and implementation of a ForCES-based router (ForTER) is illustrated. Firstly, the implementation architecture of ForTER is discussed. Then, a layered software model, which well illustrates ForCES features, is proposed. Based on the model, design and implementation of Control Element (CE) and Forwarding Element (FE) in ForTER are introduced in detail. Moreover, security for ForTER is considered and an algorithm to prevent DoS attacks is presented. Lastly, experiments of ForTER are illustrated for routing and running routing protocols, network management, DoS attack prevention, etc. The experimental results show the feasibility of the ForTER design. Consequently, the ForTER implementation basically testifies the feasibility of ForCES architecture and some IETF ForCES specifications.

Dynamic service deployment in a distributed heterogeneous cluster based router (DHCR)

This paper presents the design, implementation and evaluation of an extensible, scalable and distributed heterogeneous cluster based programmable router, called DHCR (Distributed Heterogeneous Cluster based Router), capable of supporting and deploying network services at run time. DHCR is a software IP router relying on heterogeneous cluster composed of separated computers with different hardware and software architecture capabilities, running different operating systems and interconnected through a high speed network connection. The DHCR ensures dynamic deployment of services and distributed control of router components (forwarding and routing elements) over heterogeneous system environments. The DHCR combines the IETF ForCES (Forwarding and Control Element Separation) architecture with software component technologies to meet the requirements of the next generation software routers. To ensure reliable and transparent communication between separated, decentralized and heterogeneous router components, the CORBA based middleware technology is used to support the DHCR internal communication. The paper also explores the use of the CORBA Component Model (CCM) to design and implement a modular, distributed and heterogeneous forwarding path for the DHCR router architecture. The CCM based forwarding plane ensures dynamic reconfiguration of the data path topology needed for low-level service deployment. Results on achievable performance using the proposed DHCR router are reported.

Design of programmable NoC router architecture on FPGA for multi-cluster NoCs

The next generations of massive, parallel and reconfigurable multi-cluster chips need performance and flexibility, requiring new communication schemes. According to this scenario, a programmable router architecture for networks-on-chips is presented. Results indicate that the proposed router has low area occupation, low power consumption and a high data throughput.

Active scheduling for programmable routers in open adaptive network environments

Active and Programmable networks change the functionality of intermediate nodes by using agents and active packets. This paper presents a novel packet scheduling scheme called Active Scheduling to maintain QoS in virtual private networks (VPNs) within the domain of adaptive and programmable networks. In Active Scheduling agents on the router monitor the accumulated queuing delay for each service. In order to keep the end-to-end delay within the bounds, the weights for services are adjusted dynamically by agents on the routers spanning the VPN. If there is a rise or fall in queuing delay of a service, an agent on a downstream router informs the upstream routers to alter the weights of their queues. This keeps the end-to-end delay of services within the defined bounds and offers better QoS compared to VPNs using static WFQ. The paper describes Active Scheduling, and presents simulation results and these are compared with WFQ.

A highly flexible service composition framework for real-life networks

This article introduces a novel approach for service composition on active and programmable routers. The proposed composition framework enables flexible programmability of a router's data path through dynamically loadable software components, called 'active components'. The composition framework promotes transparent, dynamic and incremental deployment of network-side functionality which renders active network technologies capable of fully replacing the existing routing infrastructure by supporting all the legacy network protocols and services, and the integration of novel ones. On the other hand, the composition framework provides the necessary structures and mechanisms to deal with services composed by independent users potentially unaware of each other. The latter is particularly important as network routers are regarded as shared resources. In order to evaluate the proposed service composition framework, we have developed an implementation for the LARA++ active router architecture [S. Schmid, A component-based active router architecture, Ph.D. Thesis, Lancaster University, UK, November 2002] and deployed several of these active routers in a real-world service network. The evaluation demonstrates the flexibility of the composition framework by enabling the alteration of service composites at run-time and provides scalability in the generation of such composites by users or administrators. This work also reports on lessons learned while deploying several active services in a production network. Besides a qualitative evaluation of the service composition framework and the LARA++ architecture, we provide a number of quantitative results, which show that the overhead of our composition model does not adversely affect the performance of the router, despite the increased flexibility achieved.