Limiting the overhead of frequent events on the control plane is essential for realizing a scalable Software-Defined Network. One way of limiting this overhead is to process frequent events in the data plane. This requires modifying switches and comes at the cost of visibility in the control plane. Taking an alternative route, we propose Kandoo, a framework for preserving scalability without changing switches. Kandoo has two layers of controllers: (i) the bottom layer is a group of controllers with no interconnection, and no knowledge of the network-wide state, and (ii) the top layer is a logically centralized controller that maintains the network-wide state. Controllers at the bottom layer run only local control applications (i.e., applications that can function using the state of a single switch) near datapaths. These controllers handle most of the frequent events and effectively shield the top layer. Kandoo's design enables network operators to replicate local controllers on demand and relieve the load on the top layer, which is the only potential bottleneck in terms of scalability. Our evaluations show that a network controlled by Kandoo has an order of magnitude lower control channel consumption compared to normal OpenFlow networks.

/pdf/kandoo-a-framework-for-efficient-and-scalable-offloading-of-2ebg6asygw.pdf

Kandoo: a framework for efficient and scalable offloading of control applications

In many distributed-memory parallel computers and high-speed communication networks, the exact structure of the underlying communication network may be ignored. These systems assume that the network creates a complete communication graph between the processors, in which passing messages is associated with communication latencies. In this paper we explore the impact of communication latencies on the design of broadcasting algorithms for fully connected message-passing systems. For this purpose, we introduce thepostal model that incorporates a communication latency parameter λ ≥ 1. This parameter measures the inverse of the ratio between the time it takes an originator of a message to send the message and the time that passes until the recipient of the message receives it. We present an optimal algorithm for broadcasting one message in systems withn processors and communication latency λ, the running time of which is Θ((λ logn)/log(λ + 1)). For broadcastingm ≥ 1 messages, we first examine several generalizations of the algorithm for broadcasting one message and then analyze a family of broadcasting algorithms based on degree-d trees. All the algorithms described in this paper are practical event-driven algorithms that preserve the order of messages.

Designing broadcasting algorithms in the postal model for message-passing systems

This paper describes our experiences, from a software perspective, with the OSIRIS network adaptor. It first identifies the problems we encountered while programming OSIRIS and optimizing network performance, and outlines how we either addressed them in the software, or had to modify the hardware. It then describes the opportunities provided by OSIRIS that we were able to exploit in the host operating system (OS); opportunities that suggested techniques for making the OS more effective in delivering network data to application programs. The most novel of these techniques, called application device channels, gives application programs running in user space direct access to the adaptor. The paper concludes with the lessons drawn from this work, which we believe will benefit the designers of future network adaptors.

/pdf/experiences-with-a-high-speed-network-adaptor-a-software-49clacoc9r.pdf

Experiences with a high-speed network adaptor: a software perspective

This paper describes a new methodology for protocol design, using incremental construction of the protocol from elements called "protocol boosters" on an as-needed basis. Protocol boosters allow: (1) dynamic protocol customization to heterogeneous environments and (2) rapid protocol evolution. Unlike alternative adaptation approaches, such as link layer, conversion, and termination protocols, protocol boosters are both robust (end-to-end protocol messages are not modified) and efficient (do not replicate the functionality of the end-to-end protocol). We give examples of error and congestion control boosters, and give initial results from booster implementations.

Protocol boosters

A successful hardware/software architecture that resolves performance bottlenecks at the workstation-to-network host interface and offers high end-to-end performance is described. The solution reported carefully splits protocol processing functions into hardware and software implementations. The interface hardware is highly parallel and performs all per-cell functions with dedicated logic to maximize performance. Software provides support for the transfer of data between the interface and application memory, as well as the state management necessary for virtual circuit setup and maintenance. In addition, all higher-level protocol processing is implemented with host software. The prototype connects a RISC System/6000 to a SONET-based asynchronous transfer model (ATM) network carrying data at the OC-3c rate of 155 Mb/s. An experimental evaluation of the interface hardware and software has been performed. Several conclusions are drawn about this host interface architecture and the workstations to which it is connected. >

/pdf/hardware-software-organization-of-a-high-performance-atm-4sqq6ge42b.pdf

Hardware/software organization of a high-performance ATM host interface

AURORA is one of five U.S. networking testbeds charged with exploring applications of, and technologies necessary for, networks operating at gigabit per second or higher bandwidths. The emphasis of the AURORA testbed, distinct from the other four testbeds, BLANCA, CASA, NECTAR and VISTANET, is research into the supporting technologies for gigabit networking. Like the other testbeds, AURORA itself is an experiment in collaboration, where government initiative (in the form of the Corporation for National Research Initiatives, which is funded by DARPA and the National Science Foundation) has spurred interaction among pre-existing centers of excellence in industry, academia, and government. AURORA has been charged with research into networking technologies that will underpin future high-speed networks. This paper provides an overview of the goals and methodologies employed in AURORA, and points to some preliminary results from our first year of research, ranging from analytic results to experimental prototype hardware. This paper enunciates our targets, which include new software architectures, network abstractions, and hardware technologies, as well as applications for our work.

https://repository.upenn.edu/cgi/viewcontent.cgi?article=1327&context=cis_reports

The AURORA gigabit testbed

We propose the development of a set of software technologies ("SwitchWare") which will enable rapid development and deployment of new network services. The key insight is that by making the basic network service selectable on a per user (or even per packet) basis, the need for formal standardization is eliminated. Additionally, by making the basic network service programmable, the deployment times, today constrained by capital funding limitations, are tremendously reduced (to the order of software distribution times). Finally, by constructing an advanced, robust programming environment, even the service development time can be reduced. A SwitchWare switch consists of input and output ports controlled by a software-programmable element; programs are contained in sequences of messages sent to the SwitchWare switch's input ports, which interpret the messages as programs. We call these "Switchlets". This accelerates the pace of network evolution, as evolving user needs can be immediately reflected in the network infrastructure. Immediate reconfigurability enhances the adaptability of the network infrastructure in the face of unexpected situations. We call a network built from SwitchWare switches an active network.

/pdf/switchware-accelerating-network-evolution-white-paper-4dapunukpm.pdf

SwitchWare: Accelerating Network Evolution (White Paper)

https://ir.nctu.edu.tw:443/bitstream/11536/14413/1/000226438500001.pdf

Meeting the broadband access infrastructure demands: The promise of Gigabit Ethernet

http://link.springer.com/content/pdf/10.1007/978-1-4615-3182-1_1.pdf

W. David Sincoskie

Papers

The AURORA gigabit testbed

SwitchWare: Accelerating Network Evolution (White Paper)

Meeting the broadband access infrastructure demands: The promise of Gigabit Ethernet

Aurora: An Experiment in Gigabit Network Technologies

Gigabit Networking Research at Bekore The AURORA gigabit network is a cooperative testbed connecting