Yan Ma

Bare Metal Provisioning to OpenStack Using xCAT

Cloud computing relies heavily on virtualization technologies. This also applies to OpenStack, which is currently the most popular IaaS platform; it mainly provides virtual machines to cloud end users. However, virtualization unavoidably brings some performance penalty, contrasted with bare metal provisioning. In this paper, we present our approach for extending OpenStack to support bare metal provisioning through xCAT(Extreme Cloud Administration Toolkit). As a result, the cloud platform could be deployed to provide both virtual machines and bare metal machines. This paper firstly introduces why bare metal machines are desirable in a cloud platform, then it describes OpenStack briefly and also the xCAT driver, which makes xCAT work with the rest of the OpenStack platform in order to provision bare metal machines to cloud end users. At the end, it presents a performance comparison between a virtualized and bare-metal environment.

Automated management of life cycle for future network experiment based on description language

Future network is a complex resources pool including multiple physical resources and virtual resources. Establishing experiment on future network is complicate and tedious. That achieving the automated management of future network experiments is so important. This paper brings forward the way for researching and managing the life cycle of experiment based on the description language. The description language uses the framework, which couples with a low hierarchical structure and a complete description of the network experiment. In this way, the experiment description template can be generated by this description framework accurately and completely. In reality, we can also customize and reuse network experiment by modifying the description template. The results show that this method can achieve the aim for managing the life cycle of network experiment effectively and automatically, which greatly saves time, reduces the difficulty, and implements the reusability of services.

SDViNet: A software defined virtual network management platform in IaaS cloud

A possible approach to address the ossification of the Internet is to set up multiple virtual networks on top of a single shared physical substrate, and each virtual network can be customized to some specific purposes. However, the current realization of virtualized infrastructure, more specifically, the IaaS cloud and network testbed, suffer from the limited programmability or high cost. To overcome these drawbacks, we propose SDViNet, a solution combining the network virtualization concept applied in IaaS cloud and SDN (Software Defined Network)-based networking testbed. It aims to achieve flexible programmable virtual network abstraction, extend the system at scale and underlying technology easily, and utilize the existing cloud platforms and commercial devices at low-cost. SDViNet is comprised of a centralized server, a set of distributed agents and a central OpenFlow controller, which communicates with the underlying packet delivery fabric. Our prototype implementation and case study show that it is feasible to provide a programmable virtual network in the IaaS cloud, while utilizing the virtualized infrastructure in an extensible and lost-cost way.

EmuCloud: An extension to build SDN test environment on IaaS cloud

In this paper, we propose a platform named EmuCloud, aiming to provide a test environment for SDN with high fidelity and low cost. EmuCloud is based on IaaS cloud, and the key idea of the platform is to extend the abstract networking resources to provide virtual openflow switch, and combine this new abstraction with existing service (such as computing, storage, authentication, etc.) seamlessly. Compared to the lightweight SDN emulator, EmuCloud provide more fidelity, extensibility and flexibility, while keeping the low-cost feature. Our prototype implementation and test case show that it is possible to build a network composed of virtual machines and fully controlled virtual openflow switches for testing SDN applications on IaaS Cloud, and the network can be adjusted dynamically to adapt to users requests.

A Routing Framework in Software Defined Network Environment

In this paper we propose a new design and prototype implementation of a routing framework in Software Defined Network (SDN) environment. This framework features a logically centralized control plane by which users could easily manage and monitor their network. It exposes APIs to support various routing protocols and algorithms so as to integrate Open Flow network with homogeneous or heterogeneous networks. We also discuss the motivation, design considerations and potential use cases in this paper.

Efficient scheduling with minimum tuning latency for WDM optical networks

Tuning latency and scheduling latency are two major overheads when designing medium access control (MAC) protocols for single hop passive star-coupled WDM optical network. Some schemes proposed so far has considered these two overheads, however, they cannot remove these two overheads completely, especially the overhead caused by receiver. To reduce the negative impact of tuning latency and scheduling latency, we propose one channel assignment algorithm and one message scheduling algorithm in this paper. The improvement of our proposed schemes in terms of average message delay can be observed from the results obtained by simulation.

Distributed Weighted Least Squares Scaling with Soft-Constraint for Node Localization in Wireless Sensor Networks

This paper presents a robust distributed localization approach which employs weighted least squares scaling with soft constrains. It incorporates the a priori deployment constraints, i.e., minimum and maximum node separation among 2-hop nodes, into localization as soft constraints and penalizes pairs of 2-hop nodes whose assigned coordinates violate the minimum and maximum constraints. Combined with applying consistency checking and statistical filtering of ranging measurements, we improve the location estimates compared with classical least squares scaling. For received signal strength based range measurements, extensive simulation results confirm that this localization scheme outperforms classical least squares scaling and is resilient against large ranging errors and sparse range measurements, which are common in wireless sensor network deployments.